Building decoration insulation board production traceability management system based on big data
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- XIAN UNIV OF POSTS & TELECOMM
- Filing Date
- 2026-06-15
- Publication Date
- 2026-07-14
Smart Images
Figure CN122390770A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of industrial big data analysis and building materials production management, specifically a big data-based traceability management system for building decoration insulation board production. Background Technology
[0002] Building insulation boards are core materials in building energy conservation projects. Their quality stability and service life during the production process are crucial to building safety. Traditional production management methods mostly adopt static recording and management based on batch information. By collecting key parameters of the production line in real time and conducting manual spot checks, basic traceability of the insulation board production process can be achieved, ensuring that the products leaving the factory meet national standards and engineering quality requirements. With the diversification of building service environments and the continuous improvement of safety standards, traditional production traceability and management models are gradually revealing their limitations. Existing technologies often lack in-depth analysis of production fluctuation characteristics during the production process, making it difficult to transform process fluctuations into quantifiable defect fingerprints, resulting in gaps in the evaluation of the product's intrinsic quality. In addition, in order scheduling and logistics, existing systems mostly adopt a simple first-in-first-out principle, lacking coupled analysis of the insulation board's intrinsic performance and the order's service environment. This leads to insufficient accuracy in predicting the long-term safe life of products under complex climatic conditions, and there is a general lack of dynamic monitoring and risk warning mechanisms during product service, making it difficult to effectively avoid engineering hazards caused by material degradation. Summary of the Invention
[0003] The purpose of this invention is to provide a big data-based traceability management system for the production of building decorative insulation boards, addressing the following technical problems: Existing technologies for traceability and management of building decorative insulation board production have significant shortcomings in converting production fluctuation characteristics into quantifiable defect fingerprints, and in predicting long-term safe lifespan and providing dynamic monitoring and early warning during service life based on the coupling of product intrinsic performance and order service environment. There is an urgent need to propose a big data-based traceability management system for building decorative insulation board production that can accurately extract defect feature fingerprints, predict safe lifespan by combining geographical environment parameters, and dynamically optimize order matching and service life risk early warning. The objective of this invention can be achieved through the following technical solutions: A big data-based traceability management system for building decoration insulation board production includes: The data acquisition terminal is configured to: collect production time sequence data of building decoration insulation boards in real time and generate corresponding batch identifiers, and upload the production time sequence data and the batch identifiers to the server; The server is configured to: receive the production time-series data and the batch identifier uploaded by the data acquisition terminal, extract features from the production time-series data to obtain a production fluctuation feature vector, and map the production fluctuation feature vector to a defect feature fingerprint; The batch identifier and the corresponding defect feature fingerprint are stored in the historical batch database; the geographical environment parameters of the target order sent by the scheduling management terminal are received, and the predicted safe lifespan is calculated in combination with the defect feature fingerprint; an order matching strategy is generated based on the comparison result between the predicted safe lifespan and the lifespan threshold preset based on the target order requirements. The scheduling management terminal is configured to: parse the target order to obtain the building service location information, convert the building service location information into the geographical environment parameters of the target order, send the geographical environment parameters to the server, receive the order matching strategy returned by the server, and execute logistics scheduling. The order matching strategy includes: when the predicted safe lifespan is less than the lifespan threshold preset based on the target order requirements, locking the matching right of the current batch of building decoration insulation boards corresponding to the batch identifier, and recommending the candidate batch with the longest predicted safe lifespan from the historical batch database; When the predicted safe lifespan is greater than or equal to the lifespan threshold preset based on the target order requirements, the current batch of building decoration insulation boards is allocated to the target order.
[0004] Optionally, the server includes: a fingerprint extraction module, configured to process the production time-series data using a preset anomaly detection algorithm, extract the production fluctuation feature vector, and map the production fluctuation feature vector to the defect feature fingerprint; The lifetime prediction module is configured to input the defect feature fingerprint and the geographical environment parameters into a preset multiphysics field environmental stress coupling prediction model, simulate and calculate the performance degradation curve, and determine the predicted safe lifetime based on the performance degradation curve. The intelligent scheduling module is configured to compare the predicted safe lifespan with the lifespan threshold preset based on the target order requirements, generate the order matching strategy, and send it to the scheduling management terminal.
[0005] Optionally, the production time series data includes temperature time series data, pressure time series data, and proportioning time series data; the defect feature fingerprint includes tensile strength decay sensitivity features and freeze-thaw sensitivity features; the preset anomaly detection algorithm adopts a long short-term memory autoencoder network.
[0006] Optionally, the geographical environment parameters include annual average temperature difference data and wind pressure data; the life prediction module is further configured to: use the preset multi-physics field environmental stress coupling prediction model, combined with the annual average temperature difference data and the wind pressure data, to calculate the failure probability distribution; and fuse the performance degradation curve and the failure probability distribution to calculate the predicted safe life.
[0007] Optionally, the system further includes: a service life monitoring terminal, configured to: collect real-time meteorological data of the service location of the building decoration insulation board, and upload the real meteorological data to the server; the server further includes: a dynamic tracking module, configured to: receive the real meteorological data, retrieve the corresponding insulation board batch identifier based on the service location information, retrieve the performance degradation curve corresponding to the batch identifier, dynamically correct the performance degradation curve using the real meteorological data to obtain the current performance degradation curve, and calculate the current safety margin based on the current performance degradation curve; and an early warning generation module, configured to: compare the current safety margin with an early warning threshold preset based on building safety specifications, generate corresponding monitoring instructions based on the comparison results, and send them to the service life monitoring terminal.
[0008] Optionally, the specific configuration of the early warning generation module is as follows: when the current safety margin is less than the early warning threshold preset based on building safety specifications, a high-altitude fall risk early warning instruction and inspection suggestion are generated and sent to the service life monitoring terminal; when the current safety margin is greater than or equal to the early warning threshold preset based on building safety specifications, a normal data recording instruction is generated and the continuous reception of the actual meteorological data is maintained.
[0009] Optionally, the scheduling management terminal includes: an order parsing module configured to obtain building service location information and building height information from the target order, and convert the building service location information and building height information into the geographical environment parameters; and a strategy execution module configured to receive the order matching strategy and generate a delivery scheduling order or inventory lock instruction according to the order matching strategy.
[0010] Optionally, the data acquisition terminal connects to the production line programmable logic controller via an industrial communication bus or hardwired to obtain the production timing data; the server is located in the cloud; the scheduling management terminal and the service life monitoring terminal are smartphones, tablets, laptops, or desktop computers.
[0011] Compared with the prior art, the present invention has the following beneficial effects: 1. This system collects multi-dimensional production time-series data of building decoration insulation boards in real time and generates batch identifiers, constructing a full-chain traceability mechanism from the source of production to the delivery end. This effectively solves the problems of insufficient mining of production fluctuation characteristics and lack of quality evaluation dimensions in the traditional management model. By using anomaly detection algorithms to extract features from the production process and map them into quantified defect feature fingerprints, the system can accurately assess the intrinsic quality of insulation boards. This not only realizes closed-loop management of production process and product quality, but also provides a standardized and structured data foundation for subsequent life prediction, significantly improving the depth and accuracy of production traceability. 2. This system introduces a multi-physics field environmental stress coupling prediction model and combines it with the geographical environmental parameters of the building's service location to calculate and predict the safe service life. This changes the logistics scheduling mode that relies solely on the first-in-first-out principle, achieving precise matching between the performance of the insulation board and environmental stress, effectively avoiding the service risks of the product under complex climatic conditions. By collecting actual meteorological data in real time through the service life monitoring terminal, the system dynamically corrects the performance degradation curve and generates early warning instructions based on the safety margin. It can proactively identify the risk of high-altitude detachment in the early stage of material performance degradation, thereby constructing a closed-loop management system for the entire life cycle of production traceability, life prediction, dynamic monitoring, and early warning disposal, greatly improving the long-term operational safety of building energy-saving projects. Attached Figure Description
[0012] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a structural diagram of the system of the present invention. Detailed Implementation
[0013] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.
[0014] Example 1: Please see Figure 1 A big data-based traceability management system for building decoration insulation board production includes: The data acquisition terminal is configured to: collect production time sequence data of building decoration insulation boards in real time and generate corresponding batch identifiers, and upload the production time sequence data and batch identifiers to the server; The server is configured to receive production time-series data and batch identifiers uploaded by the data acquisition terminal, extract features from the production time-series data to obtain production fluctuation feature vectors, and map the production fluctuation feature vectors to defect feature fingerprints. The batch identifier and corresponding defect feature fingerprint are stored in the historical batch database; the geographical environment parameters of the target order sent by the scheduling management terminal are received, and the predicted safe lifespan is calculated by combining the defect feature fingerprint; an order matching strategy is generated based on the comparison result between the predicted safe lifespan and the lifespan threshold preset based on the requirements of the target order. The scheduling and management terminal is configured to: parse the target order to obtain the building service location information, convert the building service location information into the geographical environment parameters of the target order, send the geographical environment parameters to the server, receive the order matching strategy returned by the server, and execute logistics scheduling. The order matching strategy includes: when the predicted safe lifespan is less than the lifespan threshold preset based on the target order requirements, locking the matching right of the current batch of building decoration insulation boards corresponding to the batch identifier, and recommending the candidate batch with the longest predicted safe lifespan from the historical batch database. When the predicted safe lifespan is greater than or equal to the lifespan threshold preset based on the target order requirements, the current batch of building decoration insulation boards is assigned to the target order.
[0015] This embodiment further illustrates that the data acquisition terminal synchronously collects multi-dimensional production time-series data at a fixed frequency through a sensor network deployed at key nodes of the production line to ensure accurate capture of fluctuation characteristics. Each key node preferably includes a raw material proportioning section, a heating and curing section, a pressing and molding section, and a cutting and sorting section. The server automatically merges the time-series data generated by the above nodes into the same batch record according to the same batch identifier. To ensure data closure between the traceability link and the life prediction link, the batch identifier is preferably generated by concatenating the production date, production line number, shift number, and batch serial number, and is continuously used as the main index during the warehousing, delivery, and service life monitoring stages. After receiving the geographical environment parameters of the target order, the server does not directly ship the goods according to the first-in-first-out inventory, but first retrieves the defect feature fingerprint corresponding to the batch identifier, and then calculates the lifespan in combination with the service environment of the target order, and outputs the scheduling results of whether to release the current batch and whether to recommend historical alternative batches. This forms a unified data processing link from production data collection to batch filing, defect fingerprint generation, lifespan prediction, order matching, and logistics scheduling. In this embodiment, the server can be deployed in a cloud-based distributed or microservice manner, but such deployment methods are only engineering means to achieve stable server operation and are not necessary technical features for defect fingerprint generation, lifespan prediction, or order matching strategy formation. The corresponding underlying operation and maintenance monitoring process does not participate in the core input, core calculation, and core output link of this invention. The server includes a fingerprint extraction module, configured to process production time-series data using a preset anomaly detection algorithm, extract production fluctuation feature vectors, and map the production fluctuation feature vectors to defect feature fingerprints. The life prediction module is configured to input defect feature fingerprints and geographic environment parameters into a preset multiphysics field environmental stress coupling prediction model, simulate and calculate the performance degradation curve, and determine the predicted safe life based on the performance degradation curve. The intelligent scheduling module is configured to compare the predicted safe lifespan with the lifespan threshold preset based on the target order requirements, generate an order matching strategy, and send it to the scheduling management terminal.
[0016] This embodiment further explains that the fingerprint extraction module denoises and smooths the collected time-series data to construct a structured input feature stream; the life prediction module couples the inherent defect fingerprint of the insulation board with the external environmental stress to establish a quantitative deduction process of performance degradation, thereby making the model parameter acquisition path and life prediction logic clear and reproducible. Meanwhile, to reduce structural noise caused by instantaneous network jitter during data processing in the server microservice platform, the server time-series topology graph construction step in this embodiment adopts a dynamic edge weight allocation mechanism based on a time decay factor; when extracting the call relationship between microservice nodes, the system simultaneously extracts the timestamp information of each call and calculates the edge weight based on the time difference. ; in, : Natural constant; : Edge weight, which decays as the time difference between the current system time and the timestamp of this call increases; t: Time difference between the current system time and the timestamp of this call; λ: Decay coefficient, which controls the rate of forgetting historical data, and its value is determined by the server system's tolerance for state lag; when the time difference... In this embodiment, when measured in minutes... The preferred value is to The constants between; This setting ensures that recent calls dominate the topology graph, effectively suppressing the interference of outdated historical data on the assessment of current abnormal states. This in-situ dynamic topology construction technique differs from conventional static network modeling. Its technical consideration is that frequent scaling up and down of microservice instances in core business transaction scenarios can lead to drastic changes in the system topology, and static models often exhibit significant state lag. By introducing a time decay mechanism, not only is the sensitivity of the topology graph to system state changes improved, but the utilization rate of system computing resources is also reduced to a certain extent, improving the stability of data processing when dealing with high-concurrency orders and high-frequency traffic bursts. Production time series data includes temperature time series data, pressure time series data, and proportioning time series data; defect feature fingerprints include tensile strength decay sensitivity features and freeze-thaw sensitivity features; the preset anomaly detection algorithm adopts a long short-term memory autoencoder network.
[0017] Furthermore, to clearly define the algorithm's computational rules and business logic, this embodiment structurally decomposes the Long Short-Term Memory (LSTM) autoencoder network: the encoder part receives structured spliced temperature, pressure, and ratio time-series data. Specifically, the data acquisition terminal generates a feature vector containing temperature, pressure, and ratio values at each synchronous sampling moment; continuous feature vectors are arranged in chronological order to form a shape... A two-dimensional matrix; where, This represents the number of consecutive sampling points within the sliding window. Corresponding to the three features of temperature, pressure, and ratio, this two-dimensional matrix serves as the input tensor for a single sample of the encoder; The potential temporal dependencies in the production process are extracted through a gating mechanism and compressed into a low-dimensional hidden state vector. The decoder uses the hidden state vector to reconstruct the original data; the algorithm calculates the residual sequence between the reconstructed data and the original data, and sets that when the residual value exceeds the dynamic baseline threshold, the residual sequence is extracted as a production fluctuation feature vector and mapped to the corresponding defect feature fingerprint. To enable direct reproduction of the process of mapping production fluctuation feature vectors to corresponding defect feature fingerprints, this embodiment further specifies the following: The server segments the temperature time series data, pressure time series data, and proportioning time series data using a fixed sliding window, preferably with each window covering 60 consecutive sampling points of the same process stage; for each window, seven fluctuation features are calculated: mean absolute residual, maximum residual, duration of exceeding threshold, temperature fluctuation amplitude, pressure fluctuation amplitude, proportioning deviation rate, and recovery time. Each feature is then normalized to the 0-1 interval according to the upper and lower limits of the historical training set, denoted as... to When the real-time value exceeds the upper limit of the training set, it is counted as 1; when it is below the lower limit, it is counted as 0, thereby avoiding the direct mixing of different units. The server maps the production fluctuation feature vector into two types of defect feature fingerprints according to preset weights: tensile strength attenuation sensitivity feature... Calculation; freeze-thaw sensitivity characteristics according to calculate; in, : Normalized characteristics of the mean absolute residual in the overall process fluctuation intensity; : The normalized characteristic of the maximum residual in the overall process fluctuation intensity; : Normalized feature of the duration of the abnormal persistence corresponding to the over-threshold duration; : Normalized characteristics of temperature fluctuation amplitude corresponding to temperature disturbance; : Normalized characteristics of pressure fluctuation amplitude corresponding to pressure disturbance; : Normalized characteristics of the proportion deviation rate corresponding to the proportion deviation; : Normalized feature of recovery time corresponding to the time required to recover to stability after an anomaly; : Tensile strength attenuation sensitivity characteristic, with a value range from 0 to 1. The larger the value, the more sensitive it is to the corresponding failure mode; Freeze-thaw sensitivity characteristic, with a value range from 0 to 1. The larger the value, the more sensitive the corresponding failure mode is. Minimum value function; To facilitate program invocation, this embodiment further includes and The sensitivity is divided into three levels: less than 0.30 is defined as low sensitivity, 0.30 to 0.70 as medium sensitivity, and greater than 0.70 as high sensitivity. The subsequent lifetime prediction module directly reads this value or level to participate in the calculation of the environmental stress acceleration factor. The dynamic baseline threshold is not a subjective human input, but is determined by the 95th percentile of the residuals of all normal batch windows during the training phase. When the production line equipment or raw material system changes, the threshold can be recalibrated according to the same rules. Through the above limitations, the composition of the production fluctuation feature vector, the normalization rules, the calculation method of the defect feature fingerprint, and the level classification are all clearly disclosed, so that this part can be directly reproduced and implemented. The geographical environment parameters include annual average temperature difference data and wind pressure data; the life prediction module is also configured to: use a preset multi-physics field environmental stress coupling prediction model, combined with annual average temperature difference data and wind pressure data, to calculate the failure probability distribution; and fuse the performance degradation curve and the failure probability distribution to calculate the predicted safe life.
[0018] Furthermore, in order to make the combined calculation and fusion process directly implementable as a program flow, this embodiment breaks down the life prediction module into the following steps: read meteorological records for at least ten consecutive years from the historical meteorological database of the building service location corresponding to the target order, calculate the annual average temperature difference data, and convert the wind pressure data from the building design parameters or regional wind load standard values; wherein, the annual average temperature difference data is the average of the difference between the highest monthly average temperature and the lowest monthly average temperature each year, and the wind pressure data is the design wind pressure after the building height correction; The server reads the defect feature fingerprint corresponding to the batch of insulation boards, maps the tensile strength attenuation sensitivity feature to the temperature difference stress acceleration coefficient, maps the freeze-thaw sensitivity feature to the freeze-thaw failure acceleration coefficient, and maps the wind pressure data to the interface peeling risk acceleration coefficient. The above three coefficients together form the environmental stress input vector. The lifetime prediction module first generates a baseline performance degradation curve based on defect feature fingerprints, and then performs discrete simulations according to time steps based on the environmental stress input vector. Within each time step, the material performance retention rate and failure probability are updated simultaneously to obtain a performance degradation sequence and failure probability sequence arranged by month or quarter. The multiphysics environmental stress coupling prediction model is a mathematical evolution model based on discrete time steps, and its structural parameters are specifically derived from the baseline performance degradation curve equation. and the evolution equation of actual performance retention rate and the reference strength attenuation coefficient involved Reference freeze-thaw attenuation coefficient Acceleration coefficient of thermal stress Freeze-thaw failure acceleration factor Acceleration factor of interface stripping risk constitute; To avoid the process of fusing performance degradation curves and failure probability distributions becoming uninterpretable, this embodiment specifies a dual-threshold judgment rule for predicting safe lifetime: for any calculation moment in the lifetime evolution process... If performance retention rate Less than the performance lower limit or failure probability Greater than the risk limit The moment when the condition is first met is then defined as the predicted safe life. ,Right now
[0019] in, The calculation point during the lifespan evolution process, used to represent the time difference between the time difference and the time of invocation mentioned earlier. and the following text indicating the current moment Distinguish between phases; Calculation time Performance retention rate; The lower limit of performance is determined jointly by the manufacturer's performance indicators and building safety standards, with 0.80 of the initial performance being preferred. Calculation time The probability of failure; Risk cap: The risk cap is determined by the statistics of historical project failures, with a preferred value of 0.05; when the project has higher safety requirements, the default value can be overridden by the scheduling and management terminal. Predicting safe lifespan is the moment when the dual threshold conditions are first met; To ensure the parameters are clearly sourced during program implementation, the acceleration coefficients in the multiphysics environmental stress coupling prediction model are not arbitrarily assigned manually. Instead, they are obtained through supervised fitting using batch fingerprints, corresponding service environments, and rework / sampling results from a historical batch database. When the effective sample size is insufficient to meet the preset minimum, system degradation is handled using a rule-based approach: the larger the annual average temperature difference, the higher the temperature difference stress acceleration coefficient becomes according to a piecewise linear rule; the higher the wind pressure data, the higher the interface peeling risk acceleration coefficient becomes according to a piecewise linear rule; and the higher the freeze-thaw frequency, the higher the freeze-thaw failure acceleration coefficient becomes according to a piecewise linear rule. Through this structured processing, the calculation path, fusion conditions, and threshold sources for the performance degradation curve and failure probability distribution are clearly defined, thus ensuring that the life prediction module can be directly coded and implemented. Furthermore, to eliminate the problem of unclear sources of freeze-thaw frequency, this embodiment clarifies that: the geographical environmental parameters still include annual average temperature difference data and wind pressure data, without additional changes to the input fields on the order side; the freeze-thaw frequency is an auxiliary calculation quantity derived by the server based on the historical meteorological database of the same service location when predicting the service life. Specifically, it is obtained by statistically analyzing the number of days in historical records where the daily maximum temperature is greater than 0 degrees Celsius and the next day's daily minimum temperature is less than 0 degrees Celsius, or by statistically analyzing the number of natural days when the daily maximum temperature and daily minimum temperature cross the freezing point, and then averaging them annually to form the baseline freeze-thaw frequency; In other words, the annual average temperature difference data and wind pressure data are explicit inputs or explicit conversion results on the order side, while the benchmark freeze-thaw frequency is an implicit environmental variable automatically extracted from historical meteorological records on the server side. Both are used together for subsequent lifetime calculation, but no new message fields are required for the scheduling and management terminal. To make the failure probability distribution The calculation of the basic variables is complete. This embodiment further specifies that: within each discrete time step, the system calculates the corresponding single-step damage increment based on the temperature difference stress acceleration coefficient, freeze-thaw failure acceleration coefficient and interface peeling risk acceleration coefficient, and adds the three together as the cumulative damage index. Based on a pre-defined cumulative damage index-failure probability mapping table or an equivalent monotonic probability mapping function, the cumulative damage index is converted into the current period's failure probability. The mapping table is obtained offline by analyzing the service life, repair records, and inspection results in the historical batch database, and is then permanently invoked during system runtime. Therefore... The basic variable sources are, in order: defect feature fingerprint, annual average temperature difference data, wind pressure data, and benchmark freeze-thaw frequency derived from the same historical meteorological record; with the above supplements, the interpretation chain between failure probability distribution, freeze-thaw failure acceleration coefficient, and freeze-thaw frequency in the life prediction part has been closed, thereby eliminating the problem of inconsistent symbols and input sources. To further enable those skilled in the art to stably implement the lifetime prediction module without additional creative effort, this embodiment provides the following additional limitations on each mapping and invocation rule; The acceleration factor of thermal stress was determined using a two-factor lookup table method: first, based on the sensitivity characteristics of tensile strength attenuation. It is divided into three levels: low sensitivity, medium sensitivity, and high sensitivity, and further divided into three levels based on the average annual temperature difference: less than 15 degrees Celsius, 15 to 25 degrees Celsius, and greater than 25 degrees Celsius; when For low sensitivity, the corresponding coefficients for the three levels are 1.00, 1.10, and 1.20 respectively. When the sensitivity is moderate, values of 1.10, 1.25, and 1.40 are used respectively. For high sensitivity, values of 1.20, 1.40, and 1.60 were used respectively. The freeze-thaw failure acceleration factor is also determined by looking up a table: first, according to the freeze-thaw sensitivity characteristics... It is divided into three levels: low sensitivity, medium sensitivity, and high sensitivity, and further divided into three levels based on the baseline freeze-thaw frequency derived from the server: less than 5 times per year, 5 to 20 times per year, and more than 20 times per year; when For low sensitivity, the corresponding coefficients for the three levels are 1.00, 1.12, and 1.25 respectively. When the sensitivity is moderate, values of 1.08, 1.25, and 1.45 are used respectively. For high sensitivity, values of 1.15, 1.40, and 1.65 were used respectively. The risk acceleration factor for interface stripping is determined separately by referring to a table based on wind pressure data: 1.00 when the design wind pressure is less than 0.50 kPa, 1.10 when it is between 0.50 and 0.80 kPa, and 1.25 when it is greater than 0.80 kPa; when the building height is greater than 54 meters, it is increased by 0.10 on the above basis. The initial conditions for the baseline performance degradation curve are fixed at a performance retention rate of 1 at the time of manufacture, and the discrete time step is preferably 1 month; the baseline curve is based on... generate; in, The baseline performance degradation curve at discrete month numbers Performance retention rate at the location; Discrete month number, used to indicate the month of service; : Represents the reference strength attenuation coefficient, the value of which is determined by the tensile strength attenuation sensitivity characteristics. The corresponding level is determined, and low sensitivity is selected. , sensitive to Highly sensitive ; : Represents the baseline freeze-thaw attenuation coefficient, whose value is determined by the freeze-thaw sensitivity characteristics. The corresponding level is determined, and low sensitivity is selected. , sensitive to Highly sensitive ; : The function to find the maximum value; After entering the service environment simulation, instead of refitting the entire curve, the decay amount of the baseline curve between two consecutive months is used as the basis. This decay amount is amplified by multiplying the aforementioned three acceleration coefficients, and then accumulated month by month to obtain the actual performance retention rate sequence. The specific discrete-time evolution equation is as follows:
[0020] in, To simulate and extrapolate to the 1st Actual performance retention rate over 1 month The baseline performance degradation curve is at the 1st The function value for the month, To simulate and extrapolate to the 1st Actual performance retention rate over 1 month The baseline performance degradation curve is at the 1st The function value for the month, , , These correspond to the aforementioned acceleration factors of temperature difference stress, freeze-thaw failure, and interface peeling risk, respectively. The cumulative damage index is accumulated using a fixed increment table: if the annual average temperature difference data falls into the aforementioned three ranges, the base values for single-step temperature difference damage are 0.003, 0.006, and 0.010, respectively; if the baseline freeze-thaw frequency falls into the aforementioned three ranges, the base values for single-step freeze-thaw damage are 0.000, 0.004, and 0.008, respectively; if the wind pressure data falls into the aforementioned three ranges, the base values for single-step interface stripping damage are 0.002, 0.004, and 0.007, respectively; each base value for single-step damage is then multiplied by the corresponding acceleration coefficient and summed to form the cumulative damage increment for the month; The calling rules for the cumulative damage index-failure probability mapping table are fixed as follows: when the cumulative damage index is less than 0.20, the failure probability is 0.01; 0.20 to 0.35, it is 0.03; 0.35 to 0.50, it is 0.05; 0.50 to 0.70, it is 0.10; and greater than 0.70, it is 0.20. When a more detailed project-specific mapping table is obtained through offline calibration, the project-specific mapping table is called first; otherwise, the above default mapping table is called. Through the above supplements, the mapping relationship between defect feature fingerprints and various acceleration coefficients, the initial conditions and update rules of the baseline performance decay curve, and the construction and calling methods of the cumulative damage index to failure probability have all been clearly disclosed. The dispatch management terminal includes: an order parsing module, configured to obtain building service location information and building height information from the target order, and convert the building service location information and building height information into geographical environment parameters; and a strategy execution module, configured to receive order matching strategies and generate delivery dispatch orders or inventory lock instructions according to the order matching strategies.
[0021] Furthermore, to ensure the reproducibility of the processing path for acquiring order information, converting it into geographic environment parameters, and generating delivery schedules and inventory lock instructions, the order parsing module preferably reads the target order according to a standard field template. The standard field template includes at least the project number, building service location information, building height information, required quantity, expected delivery time, and security level requirements; the building service location information includes at least one of the following: provincial / municipal / district information or latitude / longitude information. When an order contains both, latitude and longitude information is used as the primary location criterion, and provincial and municipal administrative division information is used for consistency verification. After completing the reading of the basic fields, the order parsing module first performs geocoding on the building's service location information, calls the regional meteorological database or the preset environmental parameter library to obtain the basic value of the annual average temperature difference and the regional benchmark wind pressure value corresponding to the service location, and then performs wind pressure correction based on the building height information, thereby forming the geographical environmental parameters sent to the server. To avoid the conversion of building height information into geographic environmental parameters becoming an uninterpretable process, this embodiment further stipulates that: when building height information is given directly in meters, the order parsing module selects the corresponding wind pressure correction coefficient according to the preset height grading rules; when building height information is given in terms of the number of floors, it is first converted to building height according to the preset floor height and then the same grading correction is performed; the preset floor height is preferably 3 meters, which is a common estimation parameter in building engineering and can ensure that environmental input that can be used for life prediction is quickly formed in the initial order review stage; When the building height is clearly specified in the project order, the floor conversion is no longer used; the preferred height classification rule is as follows: the first correction factor is used when the building height is not greater than 24 meters, the second correction factor is used when the building height is greater than 24 meters but not greater than 54 meters, and the third correction factor is used when the building height is greater than 54 meters; the correction factor for each level can be retrieved from the company's preset parameter table, or it can be calculated according to the applicable wind load design standard for the project location; The reason for adopting the grading rule is that there are significant differences in the wind pressure levels corresponding to different building heights. This difference directly affects the risk of interface peeling of the insulation board and the subsequent life prediction results, and is therefore a necessary parameter source affecting the invention's effectiveness. After receiving the order matching strategy returned by the server, the strategy execution module does not need to make manual intervention adjustments, but automatically generates business instructions according to the preset execution rules: when the order matching strategy indicates that the current batch of building decoration insulation boards is allocated to the target order, the strategy execution module generates a delivery schedule, which includes at least the order number, delivery batch identifier, delivery quantity, warehouse location, suggested loading order, transportation method, and expected delivery time. When the order matching strategy indicates that the matching right of the current batch of building decoration insulation boards is locked and alternative batches are recommended, the strategy execution module generates an inventory lock instruction and an alternative batch approval form. The inventory lock instruction includes at least the locked batch identifier, the reason for locking, the locking timestamp, and the lock validity period. The alternative batch approval form includes at least the recommended batch identifier, the corresponding predicted safe life, the inventory quantity, and the priority ranking. The priority ranking is not arbitrarily given, but is sorted first according to the predicted safe life from high to low. If two alternative batches have the same predicted safe life, they are then sorted secondarily according to the available inventory quantity and the outbound distance to optimize the system's logistics scheduling instructions and shorten the scheduling delay. To ensure the closed-loop traceability of the scheduling process, after generating a delivery schedule or inventory lock instruction, the scheduling management terminal also writes the execution result back to the historical batch database and order files on the server side, forming a complete link record of order input—environmental parameter generation—lifetime prediction—matching strategy output—scheduling execution. Through the above structured constraints, the field sources, parameter conversion rules, threshold basis, and output instruction content of the order parsing module and the strategy execution module are all clearly defined, and those skilled in the art can implement the corresponding software functions based on this without creative effort.
[0022] Example 2: The system also includes: a service life monitoring terminal, configured to: collect real-time meteorological data of the service location of the building decoration insulation board and upload the real-time meteorological data to the server; the server also includes: a dynamic tracking module, configured to receive the real-time meteorological data, retrieve the corresponding insulation board batch identifier based on the service location information, retrieve the performance degradation curve corresponding to the batch identifier, dynamically correct the performance degradation curve using the real-time meteorological data to obtain the current performance degradation curve, and calculate the current safety margin based on the current performance degradation curve; and an early warning generation module, configured to compare the current safety margin with the early warning threshold value preset based on building safety specifications, generate corresponding monitoring instructions based on the comparison results and send them to the service life monitoring terminal.
[0023] Furthermore, to ensure the reproducibility of the process of dynamically correcting and calculating the current safety margin, the actual meteorological data uploaded by the service life monitoring terminal must include at least the daily maximum temperature, daily minimum temperature, daily precipitation status, daily maximum instantaneous wind speed, and collection timestamp. After receiving the actual meteorological data, the dynamic tracking module retrieves the corresponding insulation board batch identifier based on the building service location information, building number, and installation date in the project delivery file. If multiple batches exist, they are distinguished one by one according to the installation area code, and the attenuation curve is maintained separately for each batch. The dynamic tracking module performs feature conversion on actual meteorological data according to calendar cycles to obtain actual temperature difference, actual freeze-thaw cycles, and actual wind pressure, and compares them with historical benchmark values in the geographic environment parameters input at the target order stage to form environmental deviation. The system uses the degradation rate of the original performance decay curve as a basis and introduces environmental deviation to correct the degradation rate: when the actual temperature difference is higher than the annual average temperature difference baseline value, the original temperature difference stress acceleration coefficient is dynamically adjusted using the following formula:
[0024] in, This is the corrected current temperature difference stress acceleration factor; This is the temperature stress acceleration factor corresponding to the original performance degradation curve; This is the positive temperature difference adjustment coefficient. The value range is determined by the frequency of occurrence of local historical extreme temperature differences, and the preferred value range is [value range missing]. to ; This represents the actual temperature difference; The reference temperature difference; When the actual number of freeze-thaw cycles or the actual wind pressure is higher than the benchmark value, the new values of the freeze-thaw failure acceleration factor and the interface peeling risk acceleration factor are calculated using the following formulas:
[0025]
[0026] in: : The corrected current freeze-thaw failure acceleration factor; : The freeze-thaw failure acceleration factor corresponding to the original performance degradation curve; : Freeze-thaw positive adjustment coefficient; Actual number of freeze-thaw cycles; Baseline freeze-thaw frequency; : The corrected current interface stripping risk acceleration factor; : The acceleration factor of interface peeling risk corresponding to the original performance degradation curve; : Positive wind pressure adjustment coefficient; Actual wind pressure; Reference wind pressure; If the value is lower than the benchmark, the same formula is used for calculation, and the adjustment term automatically becomes a reduction of the coefficient; where the subscripts in the above formula family... Represents the current state after dynamic correction using actual meteorological data, subscript Characterizes the original state predicted based on historical geographical environmental parameters; All adjustment results must be truncated and limited to preset upper and lower bounds. For example, in this embodiment, the lower bound of each acceleration coefficient is set to 0.80 times the original coefficient, and the upper bound is set to 2.00 times the original coefficient, in order to avoid distortion caused by abnormal sampling. The current performance degradation curve is generated using a rolling update method. That is, after receiving data for each statistical period, it continues to extrapolate based on the end point of the curve in the previous period, instead of recalculating the entire curve from the initial moment. To clarify the calculation rules for the current safety margin, this embodiment stipulates that: the dynamic tracking module at the current moment... Calculate the current performance retention rate respectively and current failure probability Then, the smaller remaining amount between the two relative thresholds is taken as the current safety margin. ,Right now
[0027] in, : The current moment; : Current performance retention rate at the current moment; The probability of failure at the current moment; The aforementioned performance lower limit; The aforementioned risk ceiling; The current safety margin is the smaller remaining amount between the performance retention rate component and the failure probability component relative to the threshold, i.e., it is calculated by using a minimum function. Output the minimum safety margin that will first reach the safety warning threshold; When any component is less than 0, it indicates that the corresponding dimension has crossed the safety boundary. The warning threshold is not set arbitrarily, but is determined based on the inspection trigger conditions in the building safety code and the company's existing maintenance experience. It is preferably set to 0.10. This means that a warning is triggered when there is only a 10% margin left relative to the safety boundary. In order to avoid loops in the algorithm logic deduction, the update cycle of the dynamic tracking module is fixed to once a day or once after receiving a full week of complete meteorological data. Each update only corrects the newly added time window. After the update is completed, the current performance degradation curve, the current safety margin and the corresponding timestamp are written back to the historical batch database for the subsequent warning generation module to call directly. Furthermore, in order to enable the present embodiment Compared with the aforementioned lifespan prediction stage Using the same calculation basis, this embodiment further specifies that the historical benchmark values obtained in the target order stage include not only the benchmark temperature difference and benchmark wind pressure, but also the benchmark freeze-thaw frequency automatically derived and archived by the server based on the historical meteorological records of the service location; The baseline freeze-thaw frequency does not require separate uploading by the scheduling management terminal, but is instead saved in the historical batch database along with the performance degradation curve corresponding to that batch; the dynamic tracking module calculates it at the current moment. At this time, the same damage accumulation logic as the life prediction module is adopted. That is, the damage increment of each period is first calculated based on the actual temperature difference, actual freeze-thaw cycles, and actual wind pressure in each statistical period between the installation date and the current time. Then, the increments are accumulated in chronological order to form the current cumulative damage index. The same cumulative damage index as the life prediction stage—failure probability mapping table or equivalent probability mapping function—is called to obtain the current failure probability. ; therefore, The sources of the basic variables can be clearly traced back to: actual meteorological data uploaded by the monitoring terminal during service life, installation date and installation area information in the project delivery archive, defect feature fingerprints stored in the historical batch database, and benchmark environmental variables generated and archived by the server in the initial prediction stage; with the above supplements, the sources of the benchmark freeze-thaw frequency and the basic variables for calculating the current failure probability in the dynamic tracking stage have been clarified, thus maintaining consistency with the variable definitions in the life prediction stage. The specific configuration of the early warning generation module is as follows: when the current safety margin is less than the early warning threshold preset based on the building safety code, a high-altitude fall risk early warning instruction and inspection suggestion are generated and sent to the service-life monitoring terminal; when the current safety margin is greater than or equal to the early warning threshold preset based on the building safety code, a normal data recording instruction is generated and the continuous reception of actual meteorological data is maintained.
[0028] Furthermore, in order to ensure the reproducibility of the triggering conditions and execution content for generating high-altitude fall risk warning instructions, generating inspection suggestions, and generating routine data record instructions, the warning generation module performs a graded judgment according to fixed rules after receiving the current safety margin output by the dynamic tracking module. The system reads the building height, installation area, actual wind pressure peak value in the most recent statistical period, actual freeze-thaw cycles, and current performance degradation curve slope corresponding to the building project; among them, the building height is used to determine the fall impact level, the installation area is used to distinguish corner areas, windward surfaces, and ordinary facades, and the current performance degradation curve slope is used to determine whether the degradation is in an accelerated phase. When the current safety margin is less than the warning threshold, the warning generation module shall write at least the following fields into the monitoring instruction: project number, building number, installation area code, corresponding insulation board batch identifier, warning timestamp, current safety margin value, triggering reason, and recommended handling level. The triggering reason is not just a single risk increase prompt, but is generated according to executable rules: if the performance retention rate component of the current safety margin reaches the critical boundary first, it is marked as material performance degradation-dominated risk; if the failure probability component of the current safety margin reaches the critical boundary first, it is marked as environmental failure probability-dominated risk; if both are close to the boundary at the same time, it is marked as a composite risk. Based on this, inspection recommendations are automatically generated according to preset rule templates: for buildings with a height exceeding the preset height threshold and installation areas located on windward sides or corner areas, priority is given to generating recommendations for on-site inspections of high-risk areas within 24 hours; for performance degradation curves with slopes increasing for multiple consecutive statistical periods, additional recommendations are made to check the bonding interface, anchoring components, and cracking of board joints; for actual freeze-thaw cycles significantly higher than historical benchmarks, additional recommendations are made to focus on checking freeze-thaw peeling and surface hollowing; the preset height threshold is preferably set at 24 meters, which is derived from commonly used grading standards in building facade fall risk management, making it easy for those skilled in the art to implement directly in conjunction with existing safety management processes; when projects adopt stricter enterprise standards, parameterized adjustments to this threshold are allowed without changing the algorithm process; On the other hand, when the current safety margin is greater than or equal to the warning threshold, the warning generation module generates a normal data recording instruction. This instruction includes at least the project number, building number, batch identifier, recording timestamp, current safety margin, environmental summary of the most recent statistical period, and a continued monitoring status flag. At this time, the system does not trigger a manual inspection work order, but maintains a continuous receiving status for actual meteorological data and continues to perform dynamic corrections according to the established update cycle. To avoid frequent start-stop of warnings due to slight fluctuations in monitoring values near the critical point, this embodiment further stipulates that the warning generation module adopts a delayed processing mechanism: a high-altitude fall risk warning instruction is officially issued only when the current safety margin is less than the warning critical value for two consecutive update cycles; for projects already in a warning state, to avoid ambiguity in the wording of the preset recovery ratio higher than the warning critical value, this embodiment explicitly records the warning critical value as... The preset recovery ratio is recorded as And define the recovery threshold for lifting the warning as ; in, Warning threshold; : Preset recovery ratio, preferably 1.20; The recovery threshold for lifting the warning; Therefore, projects already in a warning state will only be considered as having a safety margin that has recovered to at least the recovery threshold for two consecutive update cycles. Only when the warning is lifted and the system returns to normal data recording status; the preset recovery ratio is preferably 1.20, which means that the recovery judgment boundary is 20% higher than the trigger boundary, thereby reducing the risk of false alarms and repeated switching caused by short-term weather fluctuations; through the above structured constraints, the input fields, judgment logic, threshold source, output instruction content and state switching conditions of the warning generation module are clearly defined, so that the program can be directly implemented accordingly.
[0029] Example 3: The data acquisition terminal connects to the production line programmable logic controller via an industrial communication bus or hardwired to obtain production timing data; the server is located in the cloud; the scheduling management terminal and the service life monitoring terminal are smartphones, tablets, laptops or desktop computers.
[0030] Furthermore, to ensure that the system deployment method directly guides implementation, this embodiment defines the following limitations for each hardware interface and data flow rules: When using an industrial communication bus for access, the data acquisition terminal polls and reads temperature, pressure, proportion, start / stop status, and batch switching signals via Modbus transmission control protocol, process field network communication protocol, Ethernet control automation technology protocol, or industrial Ethernet protocol compatible with field programmable logic controllers (PLCs); when using hard-wired access, the data acquisition terminal acquires the standard voltage / current signals output by the PLC through analog input modules or digital input modules, and performs analog-to-digital conversion locally to form a unified data frame; the unified data frame must at least include the equipment number, sampling timestamp, process parameter name, parameter value, batch identifier, and checksum, so that the server side can perform time-series reconstruction by batch; To ensure reusability across different production lines, the data acquisition terminal maintains a local point mapping table. This table establishes a one-to-one correspondence between the programmable logic controller register address or terminal number and the field names of temperature time series data, pressure time series data, and proportioning time series data. When production line modifications cause changes in the points, only the mapping table is updated without altering the server's processing logic. When the server is set up in the cloud, it includes at least a data access service, a batch database service, a model calculation service, and a scheduling service. Each service interacts through interface calls. Among them, the data access service is responsible for receiving data from the data acquisition terminal and the service life monitoring terminal, the model calculation service is responsible for performing defect feature fingerprint generation, life prediction, and dynamic tracking, and the scheduling service is responsible for returning the order matching strategy to the scheduling management terminal. To avoid batch data loss due to upload interruptions, the data acquisition terminal sets up a local cache queue; when the network is abnormal, the data is temporarily stored in chronological order, and re-uploaded according to the timestamp after the network is restored, and the server performs deduplication writing according to the batch identifier + sampling timestamp; when the dispatch management terminal and the service life monitoring terminal are smartphones, tablets, laptops or desktop computers, their software interfaces all call the same set of cloud interfaces, only distinguished in terms of permissions: the dispatch management terminal has the permissions of order parsing, shipment scheduling and inventory locking, and the service life monitoring terminal has the permissions of meteorological data upload, early warning reception and inspection feedback; Through the above deployment method, the data sources, communication interfaces, field structures, cache retransmission, and permission boundaries between terminals are clearly defined, thereby enabling the system program development and joint debugging to be completed directly without creative additions.
[0031] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims
1. A big data-based traceability management system for the production of building decoration insulation boards, characterized in that: The system includes: The data acquisition terminal is configured to: collect production time sequence data of building decoration insulation boards in real time and generate corresponding batch identifiers, and upload the production time sequence data and the batch identifiers to the server; The server is configured to: receive the production time-series data and the batch identifier uploaded by the data acquisition terminal, extract features from the production time-series data to obtain a production fluctuation feature vector, and map the production fluctuation feature vector to a defect feature fingerprint; The batch identifier and the corresponding defect feature fingerprint are stored in the historical batch database; the geographical environment parameters of the target order sent by the scheduling management terminal are received, and the predicted safe lifespan is calculated in combination with the defect feature fingerprint; an order matching strategy is generated based on the comparison result between the predicted safe lifespan and the lifespan threshold preset based on the target order requirements. The scheduling management terminal is configured to: parse the target order to obtain the building service location information, convert the building service location information into the geographical environment parameters of the target order, send the geographical environment parameters to the server, receive the order matching strategy returned by the server, and execute logistics scheduling. The order matching strategy includes: when the predicted safe lifespan is less than the lifespan threshold preset based on the target order requirements, locking the matching right of the current batch of building decoration insulation boards corresponding to the batch identifier, and recommending the candidate batch with the longest predicted safe lifespan from the historical batch database; When the predicted safe lifespan is greater than or equal to the lifespan threshold preset based on the target order requirements, the current batch of building decoration insulation boards is allocated to the target order.
2. The big data-based traceability management system for building decoration insulation board production according to claim 1, characterized in that, The server includes a fingerprint extraction module, configured to process the production time series data using a preset anomaly detection algorithm, extract the production fluctuation feature vector, and map the production fluctuation feature vector to the defect feature fingerprint. The lifetime prediction module is configured to input the defect feature fingerprint and the geographical environment parameters into a preset multiphysics field environmental stress coupling prediction model, simulate and calculate the performance degradation curve, and determine the predicted safe lifetime based on the performance degradation curve. The intelligent scheduling module is configured to compare the predicted safe lifespan with the lifespan threshold preset based on the target order requirements, generate the order matching strategy, and send it to the scheduling management terminal.
3. The big data-based traceability management system for building decoration insulation board production according to claim 2, characterized in that, The production time series data includes temperature time series data, pressure time series data, and proportioning time series data; the defect feature fingerprint includes tensile strength decay sensitivity features and freeze-thaw sensitivity features; the preset anomaly detection algorithm adopts a long short-term memory autoencoder network.
4. The big data-based traceability management system for building decoration insulation board production according to claim 2, characterized in that, The geographical environment parameters include annual average temperature difference data and wind pressure data; the life prediction module is further configured to: use the preset multi-physics field environmental stress coupling prediction model, combined with the annual average temperature difference data and the wind pressure data, to calculate the failure probability distribution; and fuse the performance degradation curve and the failure probability distribution to calculate the predicted safe life.
5. The big data-based traceability management system for building decoration insulation board production according to claim 2, characterized in that, The system further includes: a service life monitoring terminal, configured to: collect real-time meteorological data of the service location of the building decoration insulation board, and upload the real meteorological data to the server; the server further includes: a dynamic tracking module, configured to: receive the real meteorological data, retrieve the corresponding insulation board batch identifier based on the service location information, retrieve the performance degradation curve corresponding to the batch identifier, dynamically correct the performance degradation curve using the real meteorological data to obtain the current performance degradation curve, and calculate the current safety margin based on the current performance degradation curve; and an early warning generation module, configured to: compare the current safety margin with an early warning threshold value preset based on building safety specifications, generate corresponding monitoring instructions based on the comparison results, and send them to the service life monitoring terminal.
6. The big data-based traceability management system for building decoration insulation board production according to claim 5, characterized in that, The specific configuration of the early warning generation module is as follows: when the current safety margin is less than the early warning threshold preset based on the building safety code, a high-altitude fall risk early warning instruction and inspection suggestion are generated and sent to the service life monitoring terminal. When the current safety margin is greater than or equal to the warning threshold preset based on building safety regulations, a normal data recording instruction is generated, and the system continues to receive the actual meteorological data.
7. The big data-based traceability management system for building decoration insulation board production according to claim 1, characterized in that, The scheduling management terminal includes: an order parsing module, configured to obtain building service location information and building height information from the target order, and convert the building service location information and building height information into the geographical environment parameters; and a strategy execution module, configured to receive the order matching strategy and generate a delivery scheduling order or inventory lock instruction according to the order matching strategy.
8. The big data-based traceability management system for building decoration insulation board production according to any one of claims 5-7, characterized in that, The data acquisition terminal obtains the production timing data by connecting to the production line programmable logic controller via an industrial communication bus or hardwired connection; the server is located in the cloud; the scheduling management terminal and the service life monitoring terminal are smartphones, tablets, laptops or desktop computers.