Cake production raw material traceability management system and method

Abstract

The present application relates to the technical field of traceability management, in particular to a cake production raw material traceability management system and method, the system comprises: an inquiry characteristic reading module analyzes a scattering signal to generate an identity sequence, a pollution interference extraction module calculates residue quality to generate an interference node, a cross-generation topology aggregation module splices the node to a data architecture to build a succession sequence when parameters are out of limit, a thermal equivalent calculation module obtains sterilization parameters to generate a correlation node according to temperature duration, and a traceability responsibility definition module investigates a cover lethal parameter to generate a boundary node, in the present application, the source unique identification is ensured by analyzing and coding fusion radio frequency characteristic extraction sequence, residue parameters are quantitatively captured by calculating container deviation, out-of-limit residue nodes are spliced to new batches to extend the axis and connect the hidden interference path, the strength is deduced by extracting the core thermal value, the fault position is locked and the isolation mark is generated by eliminating the fault position through reverse comparison of the lethal parameter, and the traceability record is extracted through deep penetration of the implicit operation.

Description

Technical Field

[0001] This invention relates to the field of traceability management technology, and in particular to a traceability management system and method for raw materials used in pastry production. Background Technology

[0002] Traceability management technology is a management system centered on information collection and business flow control. This field covers the entire business process from product manufacturing and processing, warehousing and logistics to final delivery. In specific operations, a unique feature code is assigned to the initial raw materials and the final finished products by introducing an identification and coding mechanism. The operation platform uses on-site sensing and reading equipment to obtain flow information, establishes a corresponding relationship between physical flow actions and underlying data records, and assists on-site personnel in performing supply chain control actions such as material scheduling and production planning, material feeding error prevention and interception, and reverse location of problematic batches. The traditional pastry production raw material traceability management system and method refers to the operational plan for recording the status of the flow path of raw and auxiliary materials such as flour, oil, and sugar in pastry processing. It usually uses paper labels with digital serial barcodes as information carriers. Workshop staff directly affix these paper labels to the surface of the outer packaging bags of various purchased raw materials. At the physical operation nodes such as raw material arrival inspection and warehousing, material requisition according to production order, weighing according to formula ratio, and baking, on-site operators use handheld laser barcode scanners to scan the barcodes on the packaging surface one by one. After the barcode scanner reads the internal digital serial information, the local management terminal will extract the original batch number of the material, supplier identification code, warehousing weighing value, and on-site operation timestamp corresponding to the sequence. Then, the above parameter fields are sent to the local host server through the workshop local area network line and written into the two-dimensional data form of the relational database for static storage in chronological order.

[0003] Currently, the traceability of raw materials in pastry production generally relies on manual scanning of externally affixed barcode carriers. This mechanical information entry mode is prone to physical media damage and detachment when faced with high-frequency operation in the workshop. Single-dimensional surface digital sequence collection cannot penetrate the deep state of material flow. Fixed-node intermittent data reporting mechanism causes dynamic information of intermediate processing links to be out of contact for a long time. The isolated static table storage structure severs the cross-interference relationship between upstream and downstream batches. The isolated flow of each process hinders the accurate reverse location of problematic batches and delays the efficiency of identifying the source of abnormal risks. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a traceability management system and method for pastry production raw materials.

[0005] On the one hand, a traceability management system for pastry production raw materials is provided, which includes:

[0006] The feature reading module explores the character overlap between the expected batch feature sequence and the collected RFID tag internal coding sequence, analyzes the backscattered signal associated with the RFID tag internal coding sequence with overlapping features, and generates the initial identity tracing sequence.

[0007] The pollution interference extraction module calculates the deviation between the actual weight of the emptied mixing tank and the initial baseline net weight of the equipment to extract the quality parameters of the preceding residue, and aggregates the quality parameters of the preceding residue with the initial identity tracing sequence to generate historical pollution interference nodes.

[0008] In the material mixing stage, when the quality parameters of the previous residue are determined to be greater than the preset equipment residual tolerance threshold, the cross-generational topology aggregation module splices the historical pollution interference nodes to the data architecture extension direction constructed by the actual feed weight of the current batch, and constructs a cross-generational succession traceability sequence.

[0009] The thermal equivalent calculation module analyzes the upstream branch of the cross-generational succession traceability sequence, calculates the actual sterilization thermal parameters based on the core temperature and internal residence time of the baking equipment, and aggregates the actual sterilization thermal parameters and the cross-generational succession traceability sequence to generate thermal traceability association nodes.

[0010] The traceability responsibility delineation module explores upwards along the thermal traceability associated nodes. When it determines that the actual sterilization thermal parameters cover the lethal parameters of pathogenic bacteria, it stops exploring, aggregates the stop location data to generate secondary pollution boundary marker nodes, and outputs raw material traceability management records.

[0011] As a further aspect of the present invention, the initial identity traceability sequence specifically includes the material warehousing timestamp, packaging material identification code, and supplier source code; the historical pollution interference nodes include equipment adhesion residue, previous product formula number, and handover shift number; the cross-generational succession traceability sequence specifically refers to the mixing container number, current auxiliary material ratio, and blade speed setting; the thermal traceability association nodes specifically include the furnace temperature zone number, baking chain position, and hot air circulation frequency; the secondary pollution boundary marker nodes include the abnormal triggering workstation, customer complaint batch number, and critical blockage marker; and the raw material traceability management record specifically refers to the responsibility attribution list, loss assessment and compensation amount, and workshop rectification receipt.

[0012] As a further aspect of the present invention, the query feature reading module includes:

[0013] The RFID tag screening submodule collects the inherent encoding sequence of the RFID tags attached to the goods during the material warehousing stage, extracts the preset expected batch feature sequence, performs a character-by-character comparison operation between the inherent encoding sequence of the RFID tag and the expected batch feature sequence, counts the number of identical character nodes between the inherent encoding sequence of the RFID tag and the expected batch feature sequence, extracts the effective sequence segments with overlapping features based on the number of identical character nodes, removes independent and irrelevant character branches, and obtains the effective tag code set.

[0014] The backscattering signal analysis submodule receives the backscattering signal fed back by the radio frequency hardware for the effective tag encoding set, extracts the discrete time-domain sampling points of the backscattering signal, performs a fast Fourier transform on the discrete time-domain sampling points to convert the time-domain amplitude sequence to the frequency domain dimension, extracts the phase offset parameter and amplitude attenuation parameter within the frequency domain dimension, aggregates the phase offset parameter and amplitude attenuation parameter to construct a multi-dimensional signal tensor, performs mean filtering on the multi-dimensional signal tensor to generate source tracing feature data;

[0015] The identity sequence generation submodule extracts the preset basic identification code, concatenates the source tracing feature data and the basic identification code in chronological order to construct a fused data string, divides the fused data string into equal-length data blocks, performs bitwise XOR processing on the equal-length data blocks, compresses the data bit width to output a fixed-length data string, disassembles the fixed-length data string to extract the prefix identifier segment and the core information segment, rearranges and encapsulates the prefix identifier segment and the core information segment according to the message structure, and establishes the initial identity source tracing sequence.

[0016] As a further aspect of the present invention, the process of counting the number of identical character nodes in the intrinsic encoding sequence of the RFID tag and the expected batch feature sequence, extracting effective sequence segments with overlapping features based on the number of identical character nodes, and eliminating independent and irrelevant character branches specifically involves:

[0017] Extract the total character length and fault tolerance coefficient of the expected batch feature sequence, multiply the total character length and the fault tolerance coefficient to calculate the overlap comparison threshold, perform a sliding window bit-by-bit comparison between the RFID tag's internal encoding sequence and the expected batch feature sequence, count the number of consecutively overlapping identical character nodes within the window, and extract consecutively overlapping strings when the number of identical character nodes is greater than the overlap comparison threshold. Use the consecutively overlapping strings as the valid sequence segment and delete the independent characters that do not overlap in the RFID tag's internal encoding sequence.

[0018] The process of dividing the fused data string into equal-length data blocks and performing bitwise XOR processing on the equal-length data blocks is as follows:

[0019] Read the standard bit width value defined by the communication protocol, decompose the fused data string into multiple equal-length data blocks according to the standard bit width value, extract the first equal-length data block as the initial state block, read the equal-length data blocks at subsequent sequence positions in sequence, perform a rolling XOR operation between the initial state block and the currently read equal-length data block, and output the pre-processing result.

[0020] As a further aspect of the present invention, the contamination interference extraction module includes:

[0021] The equipment weight acquisition submodule detects the emptying command trigger status of the equipment operation panel during the equipment emptying phase, drives the bottom pressure sensor to collect the pressure record of the mixing container, subtracts the benchmark value of the external support component from the pressure record of the mixing container to obtain the actual weight of the mixing cylinder after emptying, reads the factory specification nameplate to extract the initial benchmark net weight of the equipment in the absolute empty state, and aggregates the actual weight of the mixing cylinder after emptying and the initial benchmark net weight of the equipment to construct the equipment weight benchmark record;

[0022] The residue quality calculation submodule calls the equipment weight benchmark record, performs a difference calculation between the actual weight after the mixing tank is emptied and the initial benchmark net weight of the equipment to extract the primary weight deviation, performs multiple continuous sampling to calculate the average value of the primary weight deviation, filters mechanical vibration interference data to obtain stable difference data, and generates the previous residue quality parameters based on the stable difference data.

[0023] The pollution node generation submodule calls the initial identity tracing sequence and the preceding residue quality parameters, converts the preceding residue quality parameters into standardized floating-point data fields, inserts the standardized floating-point data fields as associated leaf elements into the tail identifier position of the initial identity tracing sequence, and encapsulates multi-source data messages in chronological order to construct historical pollution interference nodes.

[0024] As a further aspect of the present invention, the cross-generational topology aggregation module includes:

[0025] The residual parameter extraction submodule, during the material mixing stage, parses the data message structure contained in the historical contamination interference node, extracts the internally encapsulated preceding residue quality parameters, drives the weighing sensor array to collect the pressure data records of the mixing operation table, performs a tare weight deduction operation on the pressure data records to obtain the actual feed weight of the current batch, and aggregates the preceding residue quality parameters and the actual feed weight of the current batch to establish a feed quality reference record;

[0026] The topology splicing submodule calls the feeding quality reference record, extracts the preset equipment residual allowable error threshold, and when it is determined that the previous residue quality parameter is greater than the equipment residual allowable error threshold, it extracts the historical pollution interference node and connects it as the previous associated node to the underlying data cluster to which the actual feeding weight of the current batch belongs, and obtains the hybrid topology data architecture.

[0027] The succession sequence construction submodule extracts the node connection hierarchy within the hybrid topology data architecture, traverses downwards along the node connection hierarchy to read the associated identifier fields contained in each level, and combines multiple associated identifier fields in a unidirectional increasing order of time into line segments. Blank and faulty segments in the line segmentation process are removed to construct a continuous relationship chain and generate a cross-generational succession tracing sequence.

[0028] As a further aspect of the present invention, the process of obtaining the preset equipment residual tolerance error threshold specifically includes:

[0029] Read the equipment's factory parameters to extract the equipment's full-load design weight and weighing static drift coefficient. Multiply the equipment's full-load design weight and the weighing static drift coefficient to obtain the basic tolerance value. Collect a continuous weight signal sequence of the equipment under no-load conditions. Calculate the root mean square value of the background noise for the continuous weight signal sequence. Add the basic tolerance value and the root mean square value to generate the preset equipment residual tolerance threshold.

[0030] The process of concatenating multiple associated identifier fields in a unidirectional ascending order of time into linear segments, and eliminating blank or faulty segments in the concatenation process to construct a continuous relationship chain, specifically involves:

[0031] Extract the timestamps of the multiple associated identifier fields, connect the multiple associated identifier fields with the timestamps as indexes to construct a unidirectional time linked list, traverse the unidirectional time linked list to calculate the time span difference between adjacent nodes, parse the production scheduling plan to obtain the standard batch production cycle value, determine the fault position when the time span difference is greater than the standard batch production cycle value, and perform pointer redirection and direct connection operation on the valid nodes at both ends of the fault position to construct the continuous relationship chain.

[0032] As a further aspect of the present invention, the thermal equivalent calculation module includes:

[0033] The baking data acquisition submodule extracts the upstream branch identifier field associated with the cross-generational succession tracing sequence during the baking stage, locates the corresponding baking equipment based on the upstream branch identifier field, drives the thermocouple sensor to collect the core temperature of the baking equipment inside the baking cavity, reads the material movement trajectory recorded by the conveyor belt encoder, and calculates the internal residence time of the material exposed to the core temperature of the baking equipment by combining the material movement trajectory and the conveyor belt running speed. The baking equipment core temperature and internal residence time are aggregated to obtain the baking operation benchmark record.

[0034] The actual thermal calculation submodule reads the baking operation benchmark record, retrieves the preset standard lethal reference temperature and temperature sensitivity constant, calculates the temperature difference between the core temperature of the baking equipment and the standard lethal reference temperature, divides the temperature difference by the temperature sensitivity constant to perform exponential conversion to extract the instantaneous lethal rate, and performs continuous multiplication and accumulation of the instantaneous lethal rate and internal residence time to obtain the thermal equivalent result. Based on the thermal equivalent result, the actual sterilization thermal parameters are generated.

[0035] The thermal node aggregation submodule converts the actual sterilization thermal parameters into standard floating-point data fields, attaches the standard floating-point data fields as supplementary attribute elements to the top-level structure of the intergenerational succession tracing sequence, and performs data packet encapsulation operation on the merged full data fields in chronological order to establish thermal tracing association nodes.

[0036] As a further aspect of the present invention, the traceability responsibility determination module includes:

[0037] The thermal parameter detection submodule extracts the actual sterilization thermal parameters contained within the thermal tracing associated nodes during the investigation phase, reads the preset pathogen lethal parameters, traverses upwards level by level along the link hierarchy of the thermal tracing associated nodes, performs a difference comparison calculation between the actual sterilization thermal parameters and the pathogen lethal parameters to extract thermal intensity difference records, and generates a thermal difference detection sequence based on the thermal intensity difference records.

[0038] The boundary marker node generation submodule extracts the thermal intensity difference records within the thermal difference detection sequence, generates a coverage determination instruction when the thermal intensity difference record is not less than zero, stops the upward detection action for the link level according to the coverage determination instruction, extracts the equipment station identifier and material batch code associated with the detection stop position, and aggregates the equipment station identifier and material batch code to establish a secondary pollution boundary marker node.

[0039] The traceability record output submodule extracts the equipment station identifier and material batch code inside the secondary pollution boundary node, retrieves the finished product destination details corresponding to the equipment station identifier and material batch code along the link hierarchy, aggregates the equipment station identifier, material batch code and finished product destination details to construct a multi-dimensional responsibility traceability data table, performs formatting and encapsulation processing on the multi-dimensional responsibility traceability data table, and outputs the raw material traceability management record.

[0040] On the other hand, the method for tracing the source of raw materials for pastry production, which is based on the aforementioned traceability management system for raw materials for pastry production, includes the following steps:

[0041] S1: Compare the character overlap between the expected batch feature sequence and the collected RFID tag intrinsic coding sequence, analyze the backscattered signal associated with the RFID tag intrinsic coding sequence with overlap features, and generate the initial identity tracing sequence.

[0042] S2: Calculate the deviation between the actual weight of the emptied mixing tank and the initial baseline net weight of the equipment, extract the mass parameters of the preceding residue, and aggregate the mass parameters of the preceding residue with the initial identity tracing sequence to generate historical pollution interference nodes;

[0043] S3: During the material mixing stage, when it is determined that the quality parameter of the previous residue is greater than the preset equipment residual allowable error threshold, the historical pollution interference node is spliced ​​to the data architecture extension direction constructed by the actual feed weight of the current batch to construct a cross-generational succession traceability sequence.

[0044] S4: Analyze the upstream branch of the cross-generational succession traceability sequence, calculate the actual sterilization thermal parameters based on the core temperature and internal residence time of the baking equipment, and aggregate the actual sterilization thermal parameters and the cross-generational succession traceability sequence to generate thermal traceability association nodes.

[0045] S5: Probe upwards along the thermal traceability associated node, stop probing when the actual sterilization thermal parameters cover the lethal parameters of pathogenic bacteria, aggregate the stop position data to generate a secondary pollution boundary node, and output the raw material traceability management record.

[0046] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following:

[0047] The system analyzes the inherent coding of material labels and integrates radio frequency features to extract the initial identity sequence, ensuring a unique identifier at the source. It calculates the actual physical weight deviation before and after container emptying to capture the underlying residual interference parameters. When the residual nodes exceed the set allowable error range, they are dynamically spliced ​​onto the extension axis of the subsequent new batch data architecture, closely connecting the cross-interference paths between generations of batches. It extracts the core thermal values ​​and heating time to deduce the actual sterilization equivalent intensity. It traces back along the circulation link to compare with the baseline pathogen lethal parameters, accurately locates the boundary of contamination elimination faults and generates safety isolation marks, and deeply penetrates the hidden processing operations to extract records of potential quality hazards. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a schematic diagram of the pastry production raw material traceability management system provided in an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of the system framework of the present invention;

[0051] Figure 3 This is a flowchart of the query feature reading module in this invention;

[0052] Figure 4 This is a flowchart of the contamination interference extraction module in this invention;

[0053] Figure 5 This is a flowchart of the cross-generational topology aggregation module in this invention;

[0054] Figure 6 This is a flowchart of the thermal equivalent calculation module in this invention;

[0055] Figure 7 This is a flowchart of the traceability responsibility determination module in this invention;

[0056] Figure 8 This is a flowchart of the method for managing the traceability of raw materials for pastry production provided in an embodiment of the present invention. Detailed Implementation

[0057] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0058] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0059] This invention provides a traceability management system for pastry production raw materials, such as... Figure 1-2 The diagram shown illustrates a traceability management system for raw materials used in pastry production. This system includes:

[0060] The feature reading module explores the character overlap between the expected batch feature sequence and the collected RFID tag internal coding sequence, analyzes the backscattered signal associated with the RFID tag internal coding sequence with overlapping features, and generates the initial identity tracing sequence.

[0061] The pollution interference extraction module calculates the deviation between the actual weight of the emptied mixing tank and the initial baseline net weight of the equipment to extract the quality parameters of the preceding residue, and aggregates the quality parameters of the preceding residue with the initial identity tracing sequence to generate historical pollution interference nodes.

[0062] The cross-generational topology aggregation module, during the material mixing stage, determines that when the quality parameters of the previous residue are greater than the preset equipment residual tolerance threshold, it splices the historical pollution interference nodes to the data architecture extension direction constructed by the actual feed weight of the current batch, and constructs a cross-generational succession traceability sequence.

[0063] The thermal equivalent calculation module analyzes the upstream branch of the cross-generational succession traceability sequence, calculates the actual sterilization thermal parameters based on the core temperature and internal residence time of the baking equipment, and aggregates the actual sterilization thermal parameters and the cross-generational succession traceability sequence to generate thermal traceability association nodes.

[0064] The traceability responsibility delineation module explores upwards along the thermal traceability related nodes. When it determines that the actual sterilization thermal parameters cover the lethal parameters of pathogenic bacteria, it stops exploring, aggregates the data at the stop location to generate secondary pollution boundary marker nodes, and outputs raw material traceability management records.

[0065] The initial traceability sequence specifically includes the material warehousing timestamp, packaging material identification code, and supplier source code. Historical pollution interference nodes include equipment adhesion residue, previous product formula number, and handover shift number. The cross-generational succession traceability sequence specifically refers to the mixing container number, current auxiliary material ratio, and blade speed setting. The thermal traceability related nodes specifically include the furnace temperature zone number, baking chain position, and hot air circulation frequency. The secondary pollution boundary marker nodes include the abnormal triggering workstation, customer complaint batch number, and critical blockage marker. The raw material traceability management record specifically refers to the responsibility attribution list, loss assessment and compensation amount, and workshop rectification receipt.

[0066] Specifically, such as Figure 2 , 3 As shown, the feature reading module includes:

[0067] The RFID tag screening submodule collects the inherent encoding sequence of the RFID tags attached to the goods during the material warehousing stage, extracts the preset expected batch feature sequence, performs a character-by-character comparison operation between the inherent encoding sequence of the RFID tag and the expected batch feature sequence, counts the number of identical character nodes between the inherent encoding sequence of the RFID tag and the expected batch feature sequence, extracts the effective sequence segments with overlapping features based on the number of identical character nodes, removes independent and irrelevant character branches, and obtains the effective tag code set.

[0068] The UHF RFID antenna node receives the inherent encoded sequence of the RFID tag on the material surface. This sequence is set to a default length of 128 bits. The system indexes and extracts pre-imported expected batch feature sequences from the expected batch database within its autonomous memory region and loads these sequences into a cached comparison register. A dedicated comparison logic array is configured, and a hardware-level sliding window mechanism is initiated, fixing the window size to 8 bits wide. The sliding window continuously scans to the right across the RFID tag's inherent encoded sequence in 1-bit increments. After each window shift and lock, a bitwise XOR operation is performed between the 8-bit character within the window and the starting reference substring corresponding to the expected batch feature sequence. The XOR operation output is monitored in real time. If eight consecutive bits of the output are all 0, the starting physical memory address of the current window is recorded, and it is statically marked as a node with the same character. An accumulator is configured within the register to continuously accumulate the number of marked nodes with the same character. When the accumulator value reaches or exceeds a threshold of 16, the interception controller is activated. The address pointer reads the address of the first and last identical character nodes of the same record, copies all data bits within the physical address range to an independently allocated valid data buffer, forming a valid sequence segment with overlapping characteristics. For prefix and suffix data blocks in the original input sequence located outside the aforementioned physical address range, memory space is freed and overwritten to zero instructions are executed. Unmatched independent and irrelevant character branches are physically deleted, and finally, the valid tag encoding set in the buffer is output to the next-level data processing unit via the serial bus.

[0069] The backscattered signal analysis submodule receives the backscattered signal fed back by the radio frequency hardware for the valid tag encoding set, extracts the discrete time-domain sampling points of the backscattered signal, performs a fast Fourier transform on the discrete time-domain sampling points to convert the time-domain amplitude sequence to the frequency domain dimension, extracts the phase offset parameter and amplitude attenuation parameter within the frequency domain dimension, aggregates the phase offset parameter and amplitude attenuation parameter to construct a multi-dimensional signal tensor, performs mean filtering on the multi-dimensional signal tensor to generate source tracing feature data;

[0070] The analog-to-digital converter channel is activated to capture the radio frequency backscattered signal fed back from the valid tag code set at a sampling rate of 50 megasamples per second. A 10-millisecond electromagnetic waveform segment is extracted within the signal activity period, and the analog signal is discretized to obtain 500,000 discrete time-domain sampling points. These 500,000 discrete time-domain sampling points are then transmitted in batches to the Fast Fourier Transform (FFT) engine within the digital signal processing chip. The hardware transform points are configured to be 524,288 points to meet the power-of-two zero-padding alignment requirement. The engine projects the one-dimensional time-domain amplitude sequence into a frequency domain dimension with a center frequency of 915 MHz and a bandwidth spanning from 860 MHz to 960 MHz. The complex array sequence output in the frequency domain dimension is traversed, and the real and imaginary values ​​of each frequency component are extracted. The arithmetic logic unit is invoked to calculate the arctangent function of the quotient of the imaginary and real parts, and this result is mapped to a phase offset parameter. Simultaneously, the sum of the squares of the real and imaginary parts is calculated, and the square root operation is performed on this sum to obtain the amplitude attenuation parameter for the corresponding frequency component. A three-dimensional memory matrix space of size 256 x 256 x 2 channels is allocated. The acquired phase offset parameter is filled into the first depth channel of the matrix space, and the amplitude attenuation parameter is filled into the second depth channel, forming a multidimensional signal tensor at the physical storage level. The internal spatial filtering coprocessor is activated, defining a spatial filtering kernel matrix of size 3 x 3, setting the movement step size in both row and column directions to 1. All dimensions of the multidimensional signal tensor are traversed, and the arithmetic mean of the 18 floating-point values ​​covered by the filtering kernel is calculated. This arithmetic mean is overwritten to the original numerical bits corresponding to the center coordinates of the multidimensional signal tensor, thereby weakening high-frequency abrupt interference caused by multipath interference and generating smooth and stable source-tracing feature data.

[0071] The identity sequence generation submodule extracts the preset basic identification code, concatenates the source traceability feature data and the basic identification code in chronological order to construct a fused data string, divides the fused data string into equal-length data blocks, performs bitwise XOR processing on the equal-length data blocks, compresses the data bit width to output a fixed-length data string, disassembles the fixed-length data string to extract the prefix identifier segment and the core information segment, rearranges and encapsulates the prefix identifier segment and the core information segment according to the message structure, and establishes the initial identity source traceability sequence.

[0072] A read request is initiated to the Enterprise Resource Planning (ERP) control node via the Ethernet communication interface to retrieve a 64-bit standard basic identification code, which is temporarily stored in the input buffer queue. A 320-bit contiguous address space is allocated in the memory controller. The 256-bit traceability feature data is filled into the high-order address segment, and the 64-bit basic identification code is filled into the low-order address segment. The data is physically concatenated according to the arrival order of the timestamps, forming a fused data string with a fixed total length of 320 bits. The data fragmentation manager is activated, and the fused data string is truncated according to a pre-set 32-bit fixed segmentation span, cutting it into 10 equal-length data blocks. The first 32-bit equal-length data block at the beginning of memory is extracted and written to the accumulator register as the initial state block. The hardware-level cyclic XOR logic gate array is started. Data blocks of equal length from the 2nd to the 10th are read sequentially using a shift pointer. The currently read data block is XORed bit by bit with the value in the accumulator register, and the result is written back to the accumulator register. After 9 iterations, a single 32-bit pre-state result is output. A hash compression algorithm is used to discard the high 8 redundant parity bits in this 32-bit result sequence, compressing the payload into a 24-bit fixed-length data string. A field slicing instruction is called to remove the first 8 bits of data at the address starting point of the 24-bit sequence and allocate them as a prefix identifier segment. The remaining 16 bits are allocated as the core information segment.

[0073] Table 1 Initial Identity Sequence Message Format Table

[0074] Field Name Bit width distribution Data types Load Description Header checksum 4 people Unsigned integer Synchronization start marker Prefix Identifier Segment 8-bit Character encoding Material source attribution Core Information Segment 16-bit hexadecimal string Identity encryption features Stop character at the end of the message 4 people Unsigned integer End of frame stop bit

[0075] As shown in Table 1, based on the framework of the application layer communication message protocol, the prefix identifier segment and the core information segment are injected into the corresponding table entries, and finally encapsulated and output as a standardized initial identity tracing sequence.

[0076] Specifically, such as Figure 2 , 4 As shown, the contamination interference extraction module includes:

[0077] The equipment weight acquisition submodule detects the emptying command trigger status of the equipment operation panel during the equipment emptying phase, drives the bottom pressure sensor to collect the pressure record of the mixing container, subtracts the benchmark value of the external support component from the pressure record of the mixing container to obtain the actual weight of the mixing cylinder after emptying, reads the factory specification nameplate to extract the initial benchmark net weight of the equipment in the absolute empty state, and aggregates the actual weight of the mixing cylinder after emptying and the initial benchmark net weight of the equipment to construct the equipment weight benchmark record;

[0078] The system listens for low-level hardware interrupt signals on the control bus, extracts the clear trigger level from the main control chip on the device's operation panel, and confirms that the clear command trigger state is activated when the level jumps from 0 volts to 5 volts. It then activates four piezoelectric thin-film load cells located on the bottom support structure of the mixing container, sets the sampling refresh rate to 100 Hz, continuously captures the pressure signal fluctuations for 5 seconds, converts them into digital measurements, and records them as the mixing container's pressure record. The system indexes and loads the baseline value of the external support components from the device's static configuration parameter table. This value represents the inherent static reaction force of the unloaded support under standard gravitational acceleration, set to 150 kg. The subtraction unit inside the central processing unit subtracts 150 kg from the collected mixing container pressure record value to eliminate interference from external physical components, calculating the actual weight of the mixing tank after emptying. By scanning the metal electronic nameplate with factory specifications on the device casing using a short-range wireless reader, the internal read-only memory sector is parsed to extract the initial baseline net weight of the device in its idle state, which is constantly set to 400 kg and programmed by the manufacturer. Two independent data columns are then created in a new form row in the database. The actual weight after emptying the mixing tank is entered into the first data column, and the extracted initial baseline net weight is entered into the second data column. A unique foreign key is established between the two data columns to complete the structured construction and storage of the device weight baseline record.

[0079] The residue quality calculation submodule calls the equipment weight baseline record, calculates the difference between the actual weight after the mixing tank is emptied and the initial baseline net weight of the equipment to extract the primary weight deviation, performs multiple continuous sampling to calculate the average value of the primary weight deviation, filters mechanical vibration interference data to obtain stable difference data, and generates the previous residue quality parameters based on the stable difference data.

[0080] The system directly accesses the database form using a structured query to retrieve the equipment's baseline weight record. The actual weight of the emptied mixing cylinder returned by the query is extracted to arithmetic register A, and the initial baseline net weight of the equipment is extracted to arithmetic register B. The internal arithmetic logic unit is triggered to perform arithmetic subtraction, subtracting the value in register B from the value in register A, outputting a floating-point difference with a positive or negative sign, which is temporarily stored in main memory as the primary weight deviation. The microcontroller's internal timer interrupt service routine is activated, with a trigger interval of 200 milliseconds, to perform 50 consecutive cyclic array samplings on the primary weight deviation returned by the underlying sensors, acquiring a data set containing 50 floating-point numbers. This data set is input into a median smoothing filter. The 50 values ​​are first sorted in ascending order, and the top 5 values ​​of the highest sequence and the bottom 5 values ​​of the lowest sequence are removed to eliminate extreme peaks and valleys caused by mechanical blade vibration. The remaining 40 core data points are then arithmetically accumulated, and the sum is divided by a constant 40 to calculate a stable difference data that tends towards a static distribution. The stable difference data is multiplied by an empirical correction factor of 1.05, which characterizes the water absorption and expansion rate of the raw material, to compensate for the pseudo-mass increase of the residue adhering to the cylinder wall due to the absorption of air moisture. Finally, the corrected data is written into the variable register to generate the preceding residue mass parameters.

[0081] The pollution node generation submodule calls the initial identity tracing sequence and the previous residue quality parameters, converts the previous residue quality parameters into standardized floating-point data fields, inserts the standardized floating-point data fields as associated leaf elements into the tail identifier position of the initial identity tracing sequence, and encapsulates multi-source data messages in chronological order to construct historical pollution interference nodes.

[0082] The system locates and retrieves the memory handle for the initial identity tracing sequence from the global variable pool, while simultaneously extracting the corrected preceding residual quality parameters from the variable register. It then calls a data type conversion interface library function to forcibly convert the double-precision numeric preceding residual quality parameters into a 32-bit single-precision normalized floating-point data field conforming to the IEEE 754 standard. The system parses the underlying linked storage structure of the initial identity tracing sequence, locating the control identifier node at the tail of the linked list. A new physical memory block is allocated to store the aforementioned single-precision normalized floating-point data field, encapsulating it as an associated leaf element with an independent pointer field. The pointer field address of the tail control identifier node is modified to point to the newly created associated leaf element, thus logically inserting this element at the very end of the initial identity tracing sequence. The system obtains the timestamp network data packet protocol provided by the local operating system kernel, and combines and packages the sequence body and tail elements according to the time sequence accurate to the millisecond level. It adds a message header containing the source address and destination address, as well as a cyclic redundancy check tail, and encapsulates it into a multi-source data packet in a unified format. The data packet is then written into a distributed storage array to establish and save the fixed historical pollution interference nodes.

[0083] Specifically, such as Figure 2 , 5 As shown, the cross-generational topology aggregation module includes:

[0084] The residual parameter extraction submodule, during the material mixing stage, analyzes the data message structure contained in the historical contamination interference nodes, extracts the internally encapsulated previous residue quality parameters, drives the weighing sensor array to collect the pressure data records of the mixing operation table, performs tare weight deduction operation on the pressure data records to obtain the actual feed weight of the current batch, and aggregates the previous residue quality parameters and the actual feed weight of the current batch to establish a feed quality reference record;

[0085] The system takes over the scheduling and control of the material mixing stage, loading historical contamination interference nodes distributed on storage nodes into memory for unpacking. It identifies check bits and load boundaries in multi-source data packets, locates the end of the load segment using a pre-agreed offset pointer, and reads the internally encapsulated 32-bit standardized floating-point mass parameters of the preceding residue into the local variable space. It sends a start control frame to the weighing sensor array below the mixing platform, activating and calibrating the sensor zero point. Then, it reads the 16-bit binary digital value output by the sensor's analog-to-digital converter, maps it to a physical gravity value, and generates a pressure data record. It reads the pre-recorded tare weight of the base from the sensors when the platform is in a material-free operating state, subtracts the tare weight from the pressure data record in the subtractor, performs a tare weight deduction operation, removes the platform's own weight factor, and obtains the actual weight of the current batch of material fed into the cylinder. Create a new related data view in the relational database, using the previous residue quality parameters read earlier as the basic field of the horizontal reference coordinate system, and the calculated actual feed weight of the current batch as the comparison field of the vertical reference coordinate system. By constructing a binary tree structure, bind the two sets of values ​​under the same root node, and aggregate them into a feed quality reference record that can be directly called by the subsequent logic engine.

[0086] The topology splicing submodule calls the feeding quality reference record, extracts the preset equipment residual allowable error threshold, and when it is determined that the quality parameter of the previous residue is greater than the equipment residual allowable error threshold, it extracts the historical pollution interference node and uses it as the previous associated node to access the underlying data cluster to which the actual feeding weight of the current batch belongs, so as to obtain the hybrid topology data architecture.

[0087] The system sends an acquisition command to the data bus via an internal remote procedure call interface, loading the root node address of the binary tree of the feed quality reference record into the address register. Simultaneously, it accesses the core configuration file in the root directory, reading the preset equipment residual tolerance threshold statically entered during initialization, which is limited to 1.50 kg. It extracts the preceding residual mass parameter value from the horizontal reference coordinate system of the binary tree structure and transmits it to a hardware digital comparator for comparison with 1.50 kg using inequality logic. When the determiner outputs a high-level signal confirming that the preceding residual mass parameter is greater than 1.50 kg, it triggers the abnormal topology connection contingency plan. It activates the database retrieval pointer, extracting the historical contamination interference node containing residual weight data based on the object's unique identifier. The underlying data cluster in the database where the actual feed weight of the current batch is located is used as the target access area.

[0088] Table 2. Underlying Data Cluster Topology Connection Table

[0089] Node Identifier Node hierarchy attributes Data cluster mount point Association weight allocation 10001A Top-level root node Initialize reserved area 1.0 20002B Pre-associated nodes Buffer connection pool 0.8 30003C Underlying data cluster Business logic side 0.5 40004D Leaf terminal node Data archiving area 0.2

[0090] As shown in Table 2, by modifying the foreign key index of the relation table, the historical pollution interference node is assigned as the parent preceding associated node of the target access area, establishing a one-way causal relationship between the two. Through the change of the mapping table relationship at the underlying database, the dynamic splicing in the data architecture extension direction is completed, and finally a hybrid topology data architecture covering cross-batch association information is rendered in memory.

[0091] The succession sequence construction submodule extracts the node connection hierarchy within the hybrid topology data architecture, traverses down the node connection hierarchy to read the associated identifier fields contained in each level, and combines multiple associated identifier fields in a unidirectional ascending order of time into a segmented splicing combination. Blank and faulty segments in the segmented splicing combination process are removed to construct a continuous relationship chain and generate a cross-generational succession tracing sequence.

[0092] The rendered hybrid topology data architecture is extracted from the memory buffer and loaded into the topology parsing engine. A depth-first search algorithm is invoked to traverse the tree structure within the architecture, sequentially reading the pointer addresses representing upstream and downstream relationships in each node's storage block to extract the node connection hierarchy. Following the unidirectional downward path from parent to child nodes defined by the connection hierarchy, each data level is visited sequentially. Upon entering the data domain of each level, a 32-bit association identifier field representing batch attributes and operation characteristics is extracted. A continuous linear heap memory space is allocated, and a bubble sort algorithm is used to compare the collected association identifier fields based on their internal timestamps. The fields are then concatenated and moved along the physical memory address segments in a unidirectional ascending time order, forming a one-dimensional linear concatenated sequence. An anomaly detector is activated to scan this concatenated sequence, calculating the timestamp difference between adjacent identifier fields. If the difference exceeds the preset maximum allowable time span tolerance limit of 3600 seconds, a blank gap segment is identified within that interval. The memory splicing pointer is invoked to cross the blank fault segment, and the memory address before the fault start end is directly linked to the memory address after the fault end end. Valid memory blocks are merged, invalid gaps in the time dimension are eliminated, and a continuous and uninterrupted relationship chain is constructed. The cross-generational succession tracing sequence is then output based on this relationship chain.

[0093] Specifically, such as Figure 2 , 6 As shown, the thermal equivalent calculation module includes:

[0094] The baking data acquisition submodule extracts the upstream branch identifier field associated with the cross-generational succession traceability sequence during the baking stage. Based on the upstream branch identifier field, it locates the corresponding baking equipment, drives the thermocouple sensor to collect the core temperature of the baking equipment inside the baking cavity, reads the material movement trajectory recorded by the conveyor belt encoder, and calculates the internal residence time of the material exposed to the core temperature of the baking equipment by combining the material movement trajectory and the conveyor belt running speed. It then aggregates the core temperature of the baking equipment and the internal residence time to obtain the baking operation benchmark record.

[0095] The control command message issued at the start of the baking stage is intercepted, and the cross-generational succession traceability sequence carried within is parsed. The 16-bit upstream branch identifier field reserved at the beginning of this sequence, pointing to the upstream processing section, is extracted. Using an internal LAN broadcast query protocol, this identifier field is hash-mapped and matched with the unique physical media access control address in the workshop equipment ledger database to locate the communication port of the corresponding baking equipment executing the current batch task. A long connection of the transmission control protocol is established, sending a data retrieval request to the thermocouple sensor inside the baking equipment. The microvolt-level thermoelectric potential generated when the sensor tip contacts the hot air inside the baking cavity is collected at a 1-second interval and converted into a Celsius value as the core temperature of the baking equipment. Simultaneously, a high-precision photoelectric rotary encoder installed at the end of the conveyor belt drive shaft is connected to a reader to collect pulse signals and convert them into angular displacement, recording the linear trajectory coordinates of the material within the cavity. The constant operating speed parameter of the conveyor belt, set within the programmable logic controller (PLC), is retrieved, i.e., 0.5 meters per minute. Using an arithmetic divider, the recorded movement trajectory coordinates are divided by the conveyor belt speed to calculate the physical residence time (in seconds) of the material exposed to the effective sterilization zone of the baking equipment's core temperature. A structure variable is allocated in memory to encapsulate the core temperature value and the residence time (in seconds) within the same data packet, serving as the baseline record for the baking operation.

[0096] The actual thermal calculation submodule reads the baking operation baseline record, retrieves the preset standard lethal reference temperature and temperature sensitivity constant, calculates the temperature difference between the core temperature of the baking equipment and the standard lethal reference temperature, divides the temperature difference by the temperature sensitivity constant to perform exponential conversion to extract the instantaneous lethal rate, and performs continuous multiplication and accumulation calculation with the instantaneous lethal rate and the internal residence time to obtain the thermal equivalent result. Based on the thermal equivalent result, the actual sterilization thermal parameters are generated.

[0097] The system reads the saved baking operation baseline records from the structure variable, loading the core temperature and internal residence time of the baking equipment into the high-speed computing cache. It connects to the local process parameter server, retrieves the standard lethal reference temperature parameter (fixed at 121.1 degrees Celsius) determined experimentally for various pathogenic bacteria, and simultaneously retrieves the temperature sensitivity constant from the data table, setting its value to 10.0. The floating-point unit is activated, and a subtractor is used to subtract 121.1 degrees Celsius from the collected core temperature of the baking equipment, outputting a polar temperature difference value in a floating-point format. This temperature difference value is then divided by the temperature sensitivity constant 10.0 to obtain an intermediate ratio. Subsequently, the system calls the exponential function logic array with a base of 10 from the internal mathematical library, performing a power operation with this intermediate ratio as the exponent to calculate the instantaneous lethal rate representing the sterilization efficacy at the current temperature. The calculus accumulator is activated, with a sampling interval equal to 1 second for thermocouples. The continuously acquired instantaneous lethal rate is multiplied by the time increment, and the sum is continuously accumulated along the time axis within the internal residence time range to obtain the final thermal equivalent integral sum. This thermal equivalent result is assigned to a global variable object, completing the generation of the actual sterilization thermal parameters.

[0098] The thermal node aggregation submodule converts the actual sterilization thermal parameters into standard floating-point data fields, attaches the standard floating-point data fields as supplementary attribute elements to the top-level structure of the cross-generational succession tracing sequence, and performs data packet encapsulation operation on the merged full data fields in chronological order to establish thermal tracing association nodes.

[0099] The actual sterilization thermal parameters stored in the global variable object are extracted and formatted from their underlying binary form into 64-bit double-precision floating-point data fields conforming to the IEEE standard, preventing precision truncation and loss during long-distance data transfer. The physical memory address of the highest-level root node of the cross-generational succession tracing sequence is accessed and locked in main memory. Reserved free space is extended around the root node memory structure, and the converted standard floating-point data fields are written into this extended area, appending them as supplementary attribute elements to the top-level structure. The network encapsulation stack processing program is started, retrieving the memory-mapped file containing all data fields of the merged cross-generational succession tracing sequence. Based on the network time protocol timestamp information generated by the data, the all data fields are rearranged and reorganized in chronological order from past to present.

[0100] Table 3. Fields of Thermal Source Tracing Nodes

[0101] Field attributes Data capacity Encryption mechanism Storage tier Attribute identifier 2 bytes Plaintext transmission Top root region Double-precision floating-point segment 8 bytes Masking Attribute extension area Cross-generational succession core 128 bytes Asymmetric encryption Business Logic Area Encapsulation verification tail 4 bytes Cyclic redundancy Protocol layer tail

[0102] As shown in Table 3, based on the network packet structure customized by the traceability application layer, the recombined data body is given an application layer header, a checksum, and a security layer encryption shell. The data packet encapsulation operation is performed, and finally, a thermal traceability association node with read-only attributes is established and broadcast in the distributed ledger.

[0103] Specifically, such as Figure 2 , 7 As shown, the traceability and responsibility delineation module includes:

[0104] The thermal parameter detection submodule extracts the actual sterilization thermal parameters contained within the thermal tracing associated nodes during the investigation phase, reads the preset pathogen lethal parameters, traverses upwards level by level along the link hierarchy of the thermal tracing associated nodes, performs a difference comparison calculation between the actual sterilization thermal parameters and the pathogen lethal parameters to extract thermal intensity difference records, and generates a thermal difference detection sequence based on the thermal intensity difference records.

[0105] When switching to the background investigation phase, a retrieval request is initiated for the database form of the thermal traceability related node. The 64-bit double-precision actual sterilization thermal parameters contained in the encapsulated attribute area of ​​the node are decrypted and extracted. The pathogenic bacteria lethality parameters set for the current production environment are downloaded and read from the quality control cloud server. These parameters are the thermal integral threshold constants determined by the laboratory to ensure food safety. The link tracing algorithm interpreter is started. Using the current thermal traceability related node as the reference point, it searches backward along the pointers of the tree-like data architecture to find its direct parent node and ancestor node, traversing the entire flow link level by level upward. Each time an upstream link-level node is traced back, the historical actual sterilization thermal parameters corresponding to that node are extracted, and a subtractor is used in the arithmetic logic unit to perform a difference calculation with the standard pathogenic bacteria lethality parameters. The signed floating-point value output by the difference operation is recorded, and it is stored in a high-speed buffer queue along with the level depth identifier of the current traced node level, forming a record of thermal intensity difference mapped layer by layer. After all levels with valid pointers have been traversed, all records in the buffer queue arranged in reverse chronological order are sealed. Based on this series of hierarchically organized thermal intensity difference records, a structured thermal difference detection sequence is established.

[0106] The boundary marker node generation submodule extracts the thermal intensity difference records within the thermal difference detection sequence. When the thermal intensity difference record is not less than zero, a coverage determination instruction is generated. Based on the coverage determination instruction, the upward detection action for the link level is stopped. The equipment station identifier and material batch code associated with the detection stop position are extracted. The equipment station identifier and material batch code are aggregated to establish a secondary pollution boundary marker node.

[0107] The background parsing process extracts thermal intensity difference records line by line from the structured thermal difference detection sequence, in ascending order of hierarchy. Each extracted thermal intensity difference is loaded into the comparison instruction register of the central processing unit (CPU) and subjected to a greater-than-equal-time (GFCE) operation with a zero-level reference value. When the logic gate determines that the loaded thermal intensity difference record value is not less than zero, it determines that the actual sterilization efficiency of that node meets or exceeds the standard requirements, releases a high-level interrupt signal to the bus, and generates the highest-priority coverage determination instruction. Upon receiving this coverage determination instruction, the hardware suspends the thread resources of the link tracing algorithm interpreter, halting illegal out-of-bounds actions that would otherwise continue upwards through the link hierarchy. Using the stack pointer preserved at the point of detection stop, it reverse-maps the associated memory block where the stop location is located, reading the 16-bit device station identifier character and the 32-bit material batch code string. An independent file cluster is allocated within the non-volatile memory. The device station identifier character and the material batch code number string are concatenated using character connectors. An aggregation operation is performed to establish a secondary pollution boundary node as the basis for pollution blocking judgment, and the read and write permissions of the node are locked.

[0108] The traceability record output submodule extracts the equipment station identifier and material batch code inside the secondary pollution boundary marker node, retrieves the finished product destination details corresponding to the equipment station identifier and material batch code along the link hierarchy, aggregates the equipment station identifier, material batch code and finished product destination details to construct a multi-dimensional responsibility traceability data table, performs formatting and encapsulation processing on the multi-dimensional responsibility traceability data table, and outputs the raw material traceability management record.

[0109] The system invokes the input / output interface to send a read command to the non-volatile memory for the secondary pollution boundary marker node, reading and extracting the concatenated equipment station identifier characters and material batch code numbers within the node. Based on this equipment station identifier and material batch code, a composite primary key query condition is established in the index table of the relational database. This drives the structured query language engine to retrieve all corresponding sales orders and logistics delivery ledgers along the data link hierarchy, pulling out the detailed data records of the affected finished product destinations. A multidimensional data cube view is created in the report generator, defining the equipment station identifier as the X-axis, the material batch code as the Y-axis, and the finished product destination details as the Z-axis as the fact data source. Through the perspective and slicing functions of the data cube, the data association features of the three dimensions are aggregated to construct a multidimensional responsibility traceability data table with cross-filtering capabilities. The multidimensional responsibility traceability data table is formatted and encapsulated using a table rendering engine library, removing the underlying table header, adding the company logo, applying standard Cascading Style Sheets to control layout and indentation, generating a standard portable document format file, and outputting standard raw material traceability management records.

[0110] Please see Figure 8The method for managing the traceability of raw materials in pastry production is based on the aforementioned traceability management system for raw materials in pastry production and includes the following steps:

[0111] S1: Compare the character overlap between the expected batch feature sequence and the collected RFID tag intrinsic coding sequence, analyze the backscattered signal associated with the RFID tag intrinsic coding sequence with overlap features, and generate the initial identity tracing sequence.

[0112] S2: Calculate the deviation between the actual weight of the emptied mixing tank and the initial baseline net weight of the equipment, extract the quality parameters of the previous residue, and aggregate the quality parameters of the previous residue and the initial identity traceability sequence to generate historical pollution interference nodes;

[0113] S3: During the material mixing stage, when the quality parameters of the previous residue are determined to be greater than the preset equipment residual allowable error threshold, the historical pollution interference nodes are spliced ​​to the data architecture extension direction constructed by the actual feed weight of the current batch to construct a cross-generational succession traceability sequence.

[0114] S4: Analyze the upstream branch of the cross-generational succession traceability sequence, calculate the actual sterilization thermal parameters based on the core temperature and internal residence time of the baking equipment, and aggregate the actual sterilization thermal parameters and the cross-generational succession traceability sequence to generate thermal traceability association nodes.

[0115] S5: Probe upwards along the thermal traceability associated nodes. Stop probing when the actual sterilization thermal parameters cover the lethal parameters of pathogenic bacteria. Aggregate the stop location data to generate secondary pollution boundary node and output raw material traceability management record.

[0116] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A traceability management system for raw materials used in pastry production, characterized in that, The system includes: The feature reading module explores the character overlap between the expected batch feature sequence and the collected RFID tag internal coding sequence, analyzes the backscattered signal associated with the RFID tag internal coding sequence with overlapping features, and generates the initial identity tracing sequence. The pollution interference extraction module calculates the deviation between the actual weight of the emptied mixing tank and the initial baseline net weight of the equipment to extract the quality parameters of the preceding residue, and aggregates the quality parameters of the preceding residue with the initial identity tracing sequence to generate historical pollution interference nodes. In the material mixing stage, when the quality parameters of the previous residue are determined to be greater than the preset equipment residual tolerance threshold, the cross-generational topology aggregation module splices the historical pollution interference nodes to the data architecture extension direction constructed by the actual feed weight of the current batch, and constructs a cross-generational succession traceability sequence. The thermal equivalent calculation module analyzes the upstream branch of the cross-generational succession traceability sequence, calculates the actual sterilization thermal parameters based on the core temperature and internal residence time of the baking equipment, and aggregates the actual sterilization thermal parameters and the cross-generational succession traceability sequence to generate thermal traceability association nodes. The traceability responsibility delineation module explores upwards along the thermal traceability associated nodes. When it determines that the actual sterilization thermal parameters cover the lethal parameters of pathogenic bacteria, it stops exploring, aggregates the stop location data to generate secondary pollution boundary marker nodes, and outputs raw material traceability management records.

2. The pastry production raw material traceability management system according to claim 1, characterized in that, The initial identity traceability sequence specifically includes the material warehousing timestamp, packaging material identification code, and supplier source code. The historical pollution interference nodes include equipment adhesion residue, previous product formula number, and handover shift number. The cross-generational succession traceability sequence specifically refers to the mixing container number, current auxiliary material ratio, and blade speed setting. The thermal traceability association nodes specifically include the furnace temperature zone number, baking chain position, and hot air circulation frequency. The secondary pollution boundary marker nodes include the abnormal triggering workstation, customer complaint batch number, and critical blockage marker. The raw material traceability management record specifically refers to the responsibility attribution list, loss assessment and compensation amount, and workshop rectification receipt.

3. The pastry production raw material traceability management system according to claim 1, characterized in that, The query feature reading module includes: The RFID tag screening submodule collects the inherent encoding sequence of the RFID tags attached to the goods during the material warehousing stage, extracts the preset expected batch feature sequence, performs a character-by-character comparison operation between the inherent encoding sequence of the RFID tag and the expected batch feature sequence, counts the number of identical character nodes between the inherent encoding sequence of the RFID tag and the expected batch feature sequence, extracts the effective sequence segments with overlapping features based on the number of identical character nodes, removes independent and irrelevant character branches, and obtains the effective tag code set. The backscattering signal analysis submodule receives the backscattering signal fed back by the radio frequency hardware for the effective tag encoding set, extracts the discrete time-domain sampling points of the backscattering signal, performs a fast Fourier transform on the discrete time-domain sampling points to convert the time-domain amplitude sequence to the frequency domain dimension, extracts the phase offset parameter and amplitude attenuation parameter within the frequency domain dimension, aggregates the phase offset parameter and amplitude attenuation parameter to construct a multi-dimensional signal tensor, performs mean filtering on the multi-dimensional signal tensor to generate source tracing feature data; The identity sequence generation submodule extracts the preset basic identification code, concatenates the source tracing feature data and the basic identification code in chronological order to construct a fused data string, divides the fused data string into equal-length data blocks, performs bitwise XOR processing on the equal-length data blocks, compresses the data bit width to output a fixed-length data string, disassembles the fixed-length data string to extract the prefix identifier segment and the core information segment, rearranges and encapsulates the prefix identifier segment and the core information segment according to the message structure, and establishes the initial identity source tracing sequence.

4. The pastry production raw material traceability management system according to claim 3, characterized in that, The process of counting the number of identical character nodes between the intrinsic encoding sequence of the statistical RFID tag and the expected batch feature sequence, extracting effective sequence segments with overlapping features based on the number of identical character nodes, and eliminating independent and irrelevant character branches is as follows: Extract the total character length and fault tolerance coefficient of the expected batch feature sequence, multiply the total character length and the fault tolerance coefficient to calculate the overlap comparison threshold, perform a sliding window bit-by-bit comparison between the RFID tag's internal encoding sequence and the expected batch feature sequence, count the number of consecutively overlapping identical character nodes within the window, and extract consecutively overlapping strings when the number of identical character nodes is greater than the overlap comparison threshold. Use the consecutively overlapping strings as the valid sequence segment and delete the independent characters that do not overlap in the RFID tag's internal encoding sequence. The process of dividing the fused data string into equal-length data blocks and performing bitwise XOR processing on the equal-length data blocks is as follows: Read the standard bit width value defined by the communication protocol, decompose the fused data string into multiple equal-length data blocks according to the standard bit width value, extract the first equal-length data block as the initial state block, read the equal-length data blocks at subsequent sequence positions in sequence, perform a rolling XOR operation between the initial state block and the currently read equal-length data block, and output the pre-processing result.

5. The pastry production raw material traceability management system according to claim 3, characterized in that, The contamination interference extraction module includes: The equipment weight acquisition submodule detects the emptying command trigger status of the equipment operation panel during the equipment emptying phase, drives the bottom pressure sensor to collect the pressure record of the mixing container, subtracts the benchmark value of the external support component from the pressure record of the mixing container to obtain the actual weight of the mixing cylinder after emptying, reads the factory specification nameplate to extract the initial benchmark net weight of the equipment in the absolute empty state, and aggregates the actual weight of the mixing cylinder after emptying and the initial benchmark net weight of the equipment to construct the equipment weight benchmark record; The residue quality calculation submodule calls the equipment weight benchmark record, performs a difference calculation between the actual weight after the mixing tank is emptied and the initial benchmark net weight of the equipment to extract the primary weight deviation, performs multiple continuous sampling to calculate the average value of the primary weight deviation, filters mechanical vibration interference data to obtain stable difference data, and generates the previous residue quality parameters based on the stable difference data. The pollution node generation submodule calls the initial identity tracing sequence and the preceding residue quality parameters, converts the preceding residue quality parameters into standardized floating-point data fields, inserts the standardized floating-point data fields as associated leaf elements into the tail identifier position of the initial identity tracing sequence, and encapsulates multi-source data messages in chronological order to construct historical pollution interference nodes.

6. The pastry production raw material traceability management system according to claim 5, characterized in that, The cross-generational topology aggregation module includes: The residual parameter extraction submodule, during the material mixing stage, parses the data message structure contained in the historical contamination interference node, extracts the internally encapsulated preceding residue quality parameters, drives the weighing sensor array to collect the pressure data records of the mixing operation table, performs a tare weight deduction operation on the pressure data records to obtain the actual feed weight of the current batch, and aggregates the preceding residue quality parameters and the actual feed weight of the current batch to establish a feed quality reference record; The topology splicing submodule calls the feeding quality reference record, extracts the preset equipment residual allowable error threshold, and when it is determined that the previous residue quality parameter is greater than the equipment residual allowable error threshold, it extracts the historical pollution interference node and connects it as the previous associated node to the underlying data cluster to which the actual feeding weight of the current batch belongs, and obtains the hybrid topology data architecture. The succession sequence construction submodule extracts the node connection hierarchy within the hybrid topology data architecture, traverses downwards along the node connection hierarchy to read the associated identifier fields contained in each level, and combines multiple associated identifier fields in a unidirectional increasing order of time into line segments. Blank and faulty segments in the line segmentation process are removed to construct a continuous relationship chain and generate a cross-generational succession tracing sequence.

7. The pastry production raw material traceability management system according to claim 6, characterized in that, The process of obtaining the preset device residual tolerance error threshold is as follows: Read the equipment's factory parameters to extract the equipment's full-load design weight and weighing static drift coefficient. Multiply the equipment's full-load design weight and the weighing static drift coefficient to obtain the basic tolerance value. Collect a continuous weight signal sequence of the equipment under no-load conditions. Calculate the root mean square value of the background noise for the continuous weight signal sequence. Add the basic tolerance value and the root mean square value to generate the preset equipment residual tolerance threshold. The process of concatenating multiple associated identifier fields in a unidirectional ascending order of time into linear segments, and eliminating blank or faulty segments in the concatenation process to construct a continuous relationship chain, specifically involves: Extract the timestamps of the multiple associated identifier fields, connect the multiple associated identifier fields with the timestamps as indexes to construct a unidirectional time linked list, traverse the unidirectional time linked list to calculate the time span difference between adjacent nodes, parse the production scheduling plan to obtain the standard batch production cycle value, determine the fault position when the time span difference is greater than the standard batch production cycle value, and perform pointer redirection and direct connection operation on the valid nodes at both ends of the fault position to construct the continuous relationship chain.

8. The pastry production raw material traceability management system according to claim 6, characterized in that, The thermal equivalent calculation module includes: The baking data acquisition submodule extracts the upstream branch identifier field associated with the cross-generational succession tracing sequence during the baking stage, locates the corresponding baking equipment based on the upstream branch identifier field, drives the thermocouple sensor to collect the core temperature of the baking equipment inside the baking cavity, reads the material movement trajectory recorded by the conveyor belt encoder, and calculates the internal residence time of the material exposed to the core temperature of the baking equipment by combining the material movement trajectory and the conveyor belt running speed. The baking equipment core temperature and internal residence time are aggregated to obtain the baking operation benchmark record. The actual thermal calculation submodule reads the baking operation benchmark record, retrieves the preset standard lethal reference temperature and temperature sensitivity constant, calculates the temperature difference between the core temperature of the baking equipment and the standard lethal reference temperature, divides the temperature difference by the temperature sensitivity constant to perform exponential conversion to extract the instantaneous lethal rate, and performs continuous multiplication and accumulation of the instantaneous lethal rate and internal residence time to obtain the thermal equivalent result. Based on the thermal equivalent result, the actual sterilization thermal parameters are generated. The thermal node aggregation submodule converts the actual sterilization thermal parameters into standard floating-point data fields, attaches the standard floating-point data fields as supplementary attribute elements to the top-level structure of the intergenerational succession tracing sequence, and performs data packet encapsulation operation on the merged full data fields in chronological order to establish thermal tracing association nodes.

9. The pastry production raw material traceability management system according to claim 8, characterized in that, The source tracing and responsibility delineation module includes: The thermal parameter detection submodule extracts the actual sterilization thermal parameters contained within the thermal tracing associated nodes during the investigation phase, reads the preset pathogen lethal parameters, traverses upwards level by level along the link hierarchy of the thermal tracing associated nodes, performs a difference comparison calculation between the actual sterilization thermal parameters and the pathogen lethal parameters to extract thermal intensity difference records, and generates a thermal difference detection sequence based on the thermal intensity difference records. The boundary marker node generation submodule extracts the thermal intensity difference records within the thermal difference detection sequence, generates a coverage determination instruction when the thermal intensity difference record is not less than zero, stops the upward detection action for the link level according to the coverage determination instruction, extracts the equipment station identifier and material batch code associated with the detection stop position, and aggregates the equipment station identifier and material batch code to establish a secondary pollution boundary marker node. The traceability record output submodule extracts the equipment station identifier and material batch code inside the secondary pollution boundary node, retrieves the finished product destination details corresponding to the equipment station identifier and material batch code along the link hierarchy, aggregates the equipment station identifier, material batch code and finished product destination details to construct a multi-dimensional responsibility traceability data table, performs formatting and encapsulation processing on the multi-dimensional responsibility traceability data table, and outputs the raw material traceability management record.

10. A method for traceability management of raw materials in pastry production, characterized in that, The implementation of the pastry production raw material traceability management system according to any one of claims 1-9 includes the following steps: S1: Compare the character overlap between the expected batch feature sequence and the collected RFID tag intrinsic coding sequence, analyze the backscattered signal associated with the RFID tag intrinsic coding sequence with overlap features, and generate the initial identity tracing sequence. S2: Calculate the deviation between the actual weight of the emptied mixing tank and the initial baseline net weight of the equipment, extract the mass parameters of the preceding residue, and aggregate the mass parameters of the preceding residue with the initial identity tracing sequence to generate historical pollution interference nodes; S3: During the material mixing stage, when it is determined that the quality parameter of the previous residue is greater than the preset equipment residual allowable error threshold, the historical pollution interference node is spliced ​​to the data architecture extension direction constructed by the actual feed weight of the current batch to construct a cross-generational succession traceability sequence. S4: Analyze the upstream branch of the cross-generational succession traceability sequence, calculate the actual sterilization thermal parameters based on the core temperature and internal residence time of the baking equipment, and aggregate the actual sterilization thermal parameters and the cross-generational succession traceability sequence to generate thermal traceability association nodes. S5: Probe upwards along the thermal traceability associated node, stop probing when the actual sterilization thermal parameters cover the lethal parameters of pathogenic bacteria, aggregate the stop position data to generate a secondary pollution boundary node, and output the raw material traceability management record.

Images