A fermentation status monitoring system and method for koji blocks

Abstract

This invention provides a fermentation status monitoring system and method for koji blocks, relating to the field of fermentation process monitoring technology. The system includes multiple sensor nodes, a gateway, an access point, and a server. Multiple sensor nodes are distributed at different locations within the koji block stack to collect multi-dimensional environmental parameters, including temperature, humidity, oxygen concentration, and carbon dioxide concentration, and transmit these parameters wirelessly to the gateway. The gateway performs protocol conversion and preprocessing on the received multi-dimensional environmental parameters and sends the processed data to the access point. By deploying multiple sensor nodes at different vertical layers and horizontal positions within the koji block stack, this invention forms a three-dimensional sensing network, enabling comprehensive and accurate acquisition of real environmental parameters throughout the stack. This effectively avoids the problem of single-point monitoring, significantly reducing the probability of quality accidents caused by localized overheating or mold growth.

Description

Technical Field

[0001] This invention relates to the field of fermentation process monitoring technology, and in particular to a fermentation status monitoring system and method for koji blocks. Background Technology

[0002] Koji (fermentation starter) is a core saccharification and fermentation agent in traditional brewing industries such as baijiu, vinegar, and soy sauce. Its quality directly determines the flavor, yield, and quality stability of subsequent brewed products. The fermentation of koji is a typical solid-state microbial culture process, in which the growth, metabolism, and succession of the microbial community are highly dependent on parameters such as temperature, humidity, oxygen concentration, and carbon dioxide concentration in the fermentation environment.

[0003] Traditional monitoring of the fermentation process of koji blocks relies primarily on the experience of the koji maker, who subjectively judges the fermentation status by sight, touch, and smell. This approach has the following shortcomings: First, the koji-making workshop is large, and the koji blocks are stacked in massive quantities. The temperature and humidity distribution within the stacks is extremely uneven, making it difficult for manual inspections to comprehensively and accurately obtain the true environmental parameters of each area within the stack. This can easily lead to quality accidents such as localized overheating (koji burning) or localized mold growth. Second, while some existing monitoring equipment uses single-point sensors, they can only reflect environmental parameters near the installation point and cannot comprehensively characterize the three-dimensional environmental distribution of the koji block stacks. First, existing equipment uses single-parameter over-limit alarms, lacking a quantitative assessment of the overall state of multiple parameters. Second, the fermentation process of koji blocks includes multiple stages such as mold growth, mold drying, damp heating, high-temperature heating, post-heating, and koji maintenance. The optimal temperature, humidity, and gas concentration vary significantly at different stages. Existing equipment generally uses fixed thresholds for alarms, failing to adaptively adjust the assessment criteria according to the fermentation stage, which easily leads to false alarms or missed alarms. Third, existing equipment lacks effective preprocessing of the raw data collected by sensors. In the high humidity and strong electromagnetic interference environment of the koji-making workshop, there are many data spikes and outliers. Directly using the raw data for judgment will lead to distorted assessment results. Therefore, it is necessary to provide a koji block fermentation state monitoring system and its monitoring method that can comprehensively perceive the three-dimensional environmental distribution of koji block stacks, perform comprehensive quantitative assessment of multiple parameters, and adaptively adjust the assessment criteria according to the fermentation stage to overcome the above-mentioned technical deficiencies. Summary of the Invention

[0004] The purpose of this invention is to provide a koji block fermentation status monitoring system and its monitoring method. By deploying multiple sensor nodes at different locations in the koji block stack, the system achieves three-dimensional perception of the fermentation environment. Furthermore, by setting up a status evaluation module, a stage identification module, and an anomaly alarm module in the server, the system enables quantitative evaluation of the koji block fermentation status, stage adaptation, and anomaly early warning.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: A fermentation status monitoring system for koji blocks includes multiple sensor nodes, a gateway, an access point, and a server; Multiple sensor nodes are distributed and deployed at different locations in the curved block stack to collect multi-dimensional environmental parameters, including temperature, humidity, oxygen concentration and carbon dioxide concentration, and transmit the multi-dimensional environmental parameters to the gateway wirelessly. The gateway is used to perform protocol conversion and preprocessing on the received multidimensional environmental parameters, and then send the processed data to the access point. Access points are used to forward processed data to the server; The server is used to calculate the fermentation state index based on multidimensional environmental parameters, and outputs the multidimensional environmental parameters and fermentation state index to the display screen in a preset format for display.

[0006] Preferably, the multiple sensor nodes are divided into several layers according to the vertical height direction of the curved block stack. Each layer includes at least two sensor nodes arranged at intervals along the horizontal direction. Each sensor node is equipped with a node identifier consisting of a layer number, a row number, and a column number. The node identifier corresponds one-to-one with the spatial position of the sensor node in the curved block stack. Multiple sensor nodes and the gateway form a star topology centered on the gateway, and the data frames between the multiple sensor nodes and the gateway all carry a CRC check field. The gateway sends a retransmission instruction for data frames that fail to be received, and the maximum number of retransmissions for a single frame does not exceed 3.

[0007] Preferably, the sensor node and the gateway use the LoRa wireless communication protocol for data transmission; Data transmission between the gateway and the access point is performed using Ethernet or Wi-Fi communication protocols. Data transmission between the access point and the server is performed using the TCP / IP communication protocol. The sensor node requests SNTP time calibration from the server through the gateway every 24 hours after power-on and thereafter.

[0008] Preferably, the gateway's preprocessing includes a sliding window-based 3D model. Outlier removal and exponentially weighted moving average filtering; The length of the sliding window is , The value is a positive integer between 10 and 30. The sliding window is updated by rolling according to the first-in-first-out rule, and the mean of the sampled values ​​within the sliding window is... and standard deviation Calculate using the following formulas respectively: in Values ​​range from 0 to integers, For sampled values; when At that time, Determined as an outlier and as described above It serves as a substitute value in subsequent filtering.

[0009] Preferably, the initial value of the exponentially weighted moving average filter is... Take as the system power-on before The arithmetic mean of the next valid sample values It is a positive integer between 3 and 5.

[0010] Preferably, the server includes a status assessment module, a stage identification module, and an anomaly alarm module; The state assessment module calculates the fermentation state index based on multidimensional environmental parameters according to the following formula: The normalized membership function Using Gaussian functions: The stage identification module is used to divide the fermentation process of the koji blocks into several fermentation stages based on the changing trends of fermentation time and multidimensional environmental parameters, and dynamically adjusts the weighting coefficients according to the current fermentation stage. and the optimal value The anomaly alarm module is used to detect an anomaly based on the fermentation state index. absolute value and rate of change per unit time Generate alarm information.

[0011] Preferably, the stage identification module has a preset stage parameter configuration table. The stage parameter configuration table stores fermentation time ranges, upper and lower limits of temperature, upper and lower limits of humidity, upper and lower limits of oxygen concentration, upper and lower limits of carbon dioxide concentration, and the corresponding optimal values ​​for each stage for each stage: mold growth, mold drying, damp heating, high heat, post-heating, and koji cultivation. Tolerance coefficient and weighting coefficients ; When the stage recognition module determines that the current environmental parameters meet the judgment conditions for a certain stage, the duration of the judgment conditions must also be no less than the preset anti-shake duration. Only then can the process switch to this stage. Take 15 to 60 minutes.

[0012] Preferably, the rate of change per unit time In Take 15 to 60 minutes; The abnormal alarm module is based on the fermentation status index. The alarm levels are divided into three levels based on the size and rate of change, from light to severe: yellow alert, orange alert, and red emergency alert. Alarm information of different levels is pushed through display pop-ups, SMS platforms, and pre-bound terminal devices, respectively.

[0013] A method for monitoring the fermentation status of koji blocks includes the following steps: S1. Multiple sensor nodes collect multi-dimensional environmental parameters at different locations of the stacked curved blocks according to a preset sampling period, and send the multi-dimensional environmental parameters to the gateway wirelessly. S2. The gateway performs protocol conversion and preprocessing on the received multidimensional environmental parameters, and forwards the processed data to the server sequentially through the access point. S3. The server identifies the current fermentation stage, matches the corresponding weight coefficient and optimal value, calculates the fermentation state index based on multidimensional environmental parameters, and outputs the multidimensional environmental parameters and fermentation state index to the display screen in a preset format for display.

[0014] Preferably, after step S3, the fermentation state index is also considered. Below the preset threshold or fermentation state index rate of change per unit time When the preset change rate threshold is exceeded, the server generates alarm information according to the alarm level and pushes it in a tiered manner through the display pop-up, SMS and pre-bound terminal devices respectively; at the same time, the server maps the multi-dimensional environmental parameters and fermentation state index to the stacking space of the briquette blocks shown on the display screen in the form of a three-dimensional color temperature cloud map according to the node identifier.

[0015] The beneficial effects of this invention are: 1. This invention deploys multiple sensor nodes at different vertical layers and horizontal positions in the stack of curved blocks to form a three-dimensional perception network. This network can comprehensively and accurately obtain the real environmental parameters of various parts of the stack, effectively avoiding the problem of substituting a single point for the whole area caused by single-point monitoring, and significantly reducing the probability of quality accidents caused by local overheating or local mold growth.

[0016] 2. This invention uses a state assessment module to comprehensively calculate the fermentation state index based on a weighted normalized membership function, integrating multidimensional discrete parameters into a unified quantitative index. This allows operators to intuitively grasp the current fermentation state of the koji blocks. Compared with the traditional single-parameter limit-exceeding alarm method, the assessment dimensions are more comprehensive and the results are more objective.

[0017] 3. This invention uses a stage identification module to automatically identify the current fermentation stage based on fermentation time and parameter change trends, and adjusts the weighting coefficients and optimal parameter values ​​accordingly, thereby achieving stage-adaptive evaluation of the evaluation criteria and significantly reducing false alarm and false alarm rates.

[0018] 4. This invention uses an anomaly alarm module to issue alarms not only based on the absolute value of the fermentation state index, but also based on the rate of change of the index per unit time, which can promptly detect sudden anomalies in the fermentation process and achieve early warning.

[0019] 5. This invention performs exponentially weighted moving average filtering and outlier removal on the raw sensor data through a gateway, effectively suppressing data spikes in the strong interference environment of the koji-making workshop and improving the stability and reliability of the evaluation results. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall system architecture of the present invention; Figure 2 This is a schematic diagram showing the deployment of the sensor nodes of the present invention in a stack of curved blocks; Figure 3 This is a schematic diagram of the internal module structure of the server of the present invention; Figure 4 This is a schematic diagram of the monitoring method of the present invention; Figure 5 This is a schematic diagram of the fermentation state index calculation process of the present invention. Detailed Implementation

[0021] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Preferred embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of the invention.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0023] Specific embodiments of the present invention are described below with reference to the accompanying drawings.

[0024] Example 1 See Figure 1 As shown, this embodiment provides a koji fermentation status monitoring system, which mainly includes five parts: multiple sensor nodes, gateway, access point, server and display screen. Each part is connected in sequence through a preset wireless or wired communication link.

[0025] The sensor node is the data acquisition unit of this system. Each sensor node integrates a temperature sensor, humidity sensor, oxygen concentration sensor, carbon dioxide concentration sensor, microcontroller, LoRa wireless communication module and power supply module.

[0026] In this embodiment, temperature and humidity are measured using an integrated digital temperature and humidity sensor, model SHT30, with a temperature range of −40℃ to +125℃ and an accuracy of ±0.2℃, and a humidity range of 0 to 100%RH and an accuracy of ±2%RH; oxygen concentration is measured using an electrochemical oxygen sensor with a range of 0 to 25%; carbon dioxide concentration is measured using an NDIR non-dispersive infrared sensor with a range of 0 to 5%; the microcontroller unit is an STM32L series low-power microcontroller; the LoRa module operates in the 433MHz band, with a spreading factor of SF10, a bandwidth of 125kHz, a coding rate of 4 / 5, and a transmit power of 14dBm. Under this configuration, the effective communication distance in a brick-concrete structure workshop is no less than 300m; the power module uses a 3.6V / 19Ah lithium-ion battery, with a typical battery life of no less than 18 months under a 5-minute sampling cycle.

[0027] Considering the long-term high humidity (relative humidity often exceeding 90% RH), high dust levels, and volatile organic compound pollution in the koji-making workshop, the sensor node in this embodiment adopts an injection-molded one-piece shell with an IP65 protection rating, meaning it is entirely dustproof and can withstand water spray from any direction. The gas sensor probe at the shell opening is covered with a hydrophobic and breathable polytetrafluoroethylene membrane, which ensures gas diffusion and exchange while preventing liquid moisture and dust from entering. The surface of the node shell is equipped with a snap-fit ​​structure for easy fixing, so that the node can be securely fastened to the special bracket for stacking koji blocks.

[0028] Multiple sensor nodes are divided into several layers according to the vertical height of the stacked blocks, and each layer includes at least two sensor nodes arranged at intervals along the horizontal direction.

[0029] See Figure 2 As shown, in a typical stack of curved blocks, there are four layers vertically from bottom to top. Each layer has one sensor node near each of the four corners and one at the center, meaning five sensor nodes per layer, for a total of 20 sensor nodes in the stack. Each sensor node is assigned a unique three-digit identifier, representing the layer number L, row number R, and column number C. For example, L2-R1-C3 represents the sensor node located in the first row and third column of the second layer. The node identifier corresponds one-to-one with the spatial position of the sensor node within the stack, facilitating the server-side organization, display, and analysis of data based on location.

[0030] Multiple sensor nodes and the gateway form a star topology centered on the gateway. That is, all sensor nodes communicate directly with the gateway in a single hop. When the nodes send data, they use the time-slotted ALOHA access method of LoRaWAN Class A. The transmission time is pre-allocated to different time slots according to the node identifier to reduce the probability of air collisions caused by multiple nodes transmitting at the same time in the same koji production workshop.

[0031] Each sensor node collects environmental parameters every 5 minutes. After collection, the data is assembled into a frame. The frame format includes the following in sequence: frame header (2 bytes), node identifier (3 bytes), timestamp (4 bytes), temperature (2 bytes), humidity (2 bytes), oxygen concentration (2 bytes), carbon dioxide concentration (2 bytes), battery voltage (1 byte), and CRC-16 checksum (2 bytes).

[0032] After receiving a data frame, the gateway first verifies the CRC-16 field. If the verification passes, it sends an ACK confirmation frame back to the node; otherwise, it sends a NACK retransmission command back to the node. Each sensor node starts a retransmission waiting timer (T_ack=3s) after sending a data frame. If no ACK is received before the timer expires, it automatically retransmits once, with a maximum of 3 retransmissions. If the maximum number of retransmissions is exceeded and the transmission is still unsuccessful, the data is discarded, and the packet loss event is recorded in the local Flash memory for later analysis.

[0033] To ensure the consistency of timestamps among multiple nodes, each sensor node initiates an SNTP time synchronization request to the server through the gateway upon power-on and thereafter every 24 hours. The request message carries the node's local timestamp, and the server returns the server timestamp. The node calculates and adjusts its local real-time clock based on the round-trip time difference. After adjustment, the clock deviation between the node and the server is controlled within ±1 second.

[0034] The gateway maintains a sliding window buffer of length N for each type of environmental parameter uploaded by each node. Take a positive integer between 10 and 30 in this embodiment. =20, meaning the sliding window can hold a maximum of the 20 most recent valid sample values. Each new sample value... When the sample arrives, the sliding window is updated according to the first-in-first-out rule, that is, the earliest sampled value in the window is discarded and the new sampled value is placed at the end of the window.

[0035] mean of the sampled values ​​within the sliding window and standard deviation Calculate using the following formulas respectively: when At that time, the new sampled value It was identified as an outlier and... Replace with the mean of the current sliding window The mean substitute value is used in subsequent EWMA filtering operations.

[0036] It should be noted that the replaced The outlier will not be re-injected into the sliding window to participate in the update of the mean and standard deviation, in order to prevent a single outlier from polluting subsequent statistical parameters.

[0037] For the cold start phase when the system has just started and the sliding window has not yet been filled (i.e., the number of samples received has been...) < (At that time), directly use the existing ones. Calculate the mean and standard deviation of each sample value. The temporary relaxation was determined to be To avoid insufficient samples during the cold start phase. The issue of being too small led to false rejections; once achieve Then restore the standard determination.

[0038] After removing outliers, the gateway applies an exponentially weighted moving average filter to the data, the expression of which is: in For the first The filtered output value of the next sample. For filter coefficients, 0 < <1.

[0039] This embodiment focuses on temperature and humidity parameters. =0.3, taking the oxygen and carbon dioxide concentration parameters as follows: =0.2, balancing noise immunity and response speed.

[0040] Initial value of filter The initialization rules are as follows: After each sensor node powers on, the gateway discards the first two sampled values ​​of that node (to avoid the bias during the sensor warm-up phase) and takes the subsequent consecutive values. The arithmetic mean of the next effective sample values ​​is used as , Take a positive integer between 3 and 5, in this embodiment =3. This initialization strategy can effectively avoid long-term shifts in the filter output caused by occasional glitches when using the first sampled value as the initial value.

[0041] The server hardware uses industrial-grade servers or private cloud servers, and the software deploys data receiving services, databases, and a monitoring and management platform, further including a status assessment module, a stage identification module, and an anomaly alarm module. (See [link to relevant documentation]). Figure 3 As shown.

[0042] The database uses the time-series database InfluxDB to store the raw data and filtered data, and the relational database MySQL to store node configuration information, stage parameter configuration tables, and alarm history records.

[0043] The stage identification module has a pre-set stage parameter configuration table. The stage parameter configuration table is divided into six stages according to the typical stages of the fermentation process of the koji blocks: mold growth, mold drying, damp fire, high fire, post-fire, and koji maintenance. Each stage corresponds to the storage of fermentation time range, upper and lower limits of temperature, upper and lower limits of humidity, upper and lower limits of oxygen and carbon dioxide concentration, as well as the optimal value. Tolerance coefficient and weighting coefficients For a typical configuration of this embodiment, please refer to Table 1: Table 1: Typical Configurations (Temperature in °C, Humidity in %RH, O2 and CO2 in % volume fraction) It should be noted that the four components in the weight vector w correspond to temperature, humidity, O2, and CO2, respectively, and the sum of the four is equal to 1. The above values ​​are the recommended configuration for typical baijiu daqu in this embodiment. In actual applications, they can be fine-tuned according to the specific daqu-making process.

[0044] The stage identification module determines the current fermentation stage according to the following priority rules: ① Based on the fermentation time, roughly determine the current stage; ②Accurate correction is made based on the interval assignment of the measured temperature values; ③ The humidity was further corrected based on the interval assignment of the measured humidity values; ④ Use the measured values ​​of O2 and CO2 as auxiliary verification parameters. When the judgment results of multiple parameters conflict, arbitration is carried out according to the priority of temperature > humidity > CO2 > O2, and the parameter with the larger weight is determined.

[0045] To avoid frequent jumps in stage recognition caused by instantaneous parameter fluctuations, this embodiment introduces a stabilization duration constraint: Only when the current parameter combination meets the judgment condition for a certain stage and the condition is maintained continuously for a duration not less than the anti-shake duration. At that time, the stage identification module switches to that stage. Take for 15 to 60 minutes, in this embodiment =30 minutes. In the critical region of phase switching, the module also introduces hysteresis of ±1℃ and ±2%RH for temperature and humidity to further suppress boundary jitter.

[0046] The status assessment module obtains the current stage index from the stage identification module, and then matches the corresponding index from the stage parameter configuration table. , and Calculate the fermentation state index according to the following formula. : Normalized membership function Using Gaussian functions: In this embodiment, n=4. These correspond to temperature, humidity, O2, and CO2, respectively. The value of S ranges from [0, 1]. A value closer to 1 indicates that the fermentation environment is closer to the ideal state. The state assessment module calculates the position of each sensor node. Values, and for the entire stack of 20 nodes. The stacking level is obtained by taking the arithmetic mean of the values. .

[0047] The abnormal alarm module is based on absolute value and rate of change per unit time Multi-level alarms are implemented, including Take for 15 to 60 minutes, in this embodiment =30 minutes, meaning the rate of change is calculated every 30 minutes. Alarms are categorized into three levels based on severity, from mild to severe: Yellow alert: 0.70 ≤ S < 0.85 or 0.05 ≤ If the value is less than 0.10 (every 30 minutes), a pop-up notification will appear on the display screen to remind the staff on duty to pay attention. Orange alert: 0.50 ≤ S < 0.70 or 0.10 ≤ <0.20 (every 30 minutes), in addition to the pop-up window on the display screen, will be pushed to the pre-bound workshop supervisor's mobile phone via SMS platform; Red emergency alert: <0.50 or ≥0.20 (every 30 minutes), in addition to the two methods mentioned above, further push notifications are sent to the bound terminals of production scheduling, quality control and equipment operation and maintenance positions via mobile APP, and the spatial location of the abnormal node is highlighted on the display screen in a flashing manner.

[0048] The formula for calculating the rate of change per unit time is: To avoid false alarms triggered by a single spike, the module requires that alarm conditions be met continuously for at least two sampling periods (i.e., 10 minutes) before an alarm is officially triggered; similarly, alarm cancellation requires that the cancellation conditions be met continuously for two sampling periods after the conditions are cleared before it is automatically cleared.

[0049] Alarm information is stored in a structured manner, including alarm level, occurrence time, node identifier, node spatial location, abnormal parameter name, and measured value of the abnormal parameter. value, Alarm values ​​and recommended handling measures. Alarm history records can be queried, filtered, and reviewed on the monitoring management platform.

[0050] The display screen is connected to the server via a video output interface or a network interface, and the displayed content mainly includes three parts.

[0051] The first part is a 3D color temperature cloud map of the stacked curved blocks. The server maps the measured position of each node to coordinates in the 3D visualization space according to the node identifier LRC. The mapping rules are as follows: in , , These are the display scale factors for column spacing, row spacing, and layer spacing, respectively. Each node is displayed as a spherical marker, and the color of the spherical marker is interpolated onto a green-yellow-red band according to its S value, specifically: when... Values ​​≥0.85 are displayed in green, and values ​​≤0.70 are displayed in green. <0.85 is yellow, 0.50≤ Orange when <0.70 Red indicates a value less than 0.50. Isosurfaces are generated between nodes through spatial interpolation to visually reflect the state distribution inside the stack.

[0052] The second part is the parameter details panel for the selected node, including node identifier, spatial location, current temperature, humidity, measured O2 and CO2 values, etc. Value, current fermentation stage, and duration of the stage.

[0053] The third part is the time series curve panel, which displays specified parameters or time windows selected by the user. Historical change curve of the value.

[0054] When an alarm occurs, the alarm information is displayed in a pop-up window at the top of the screen, and the corresponding node in the 3D cloud map flashes at a frequency of 1Hz for emphasis.

[0055] See Figure 4 and Figure 5As shown, the working process is as follows: Step S1: After the system starts, 20 sensor nodes synchronously collect temperature, humidity, O2 and CO2 data at their locations at a sampling period of 5 minutes. After assembling the data into a data frame with CRC-16 check, the data is sent to the gateway via the LoRa wireless communication protocol. The gateway performs CRC check and ACK / NACK feedback on each frame and retransmits it up to a maximum of 3 times if necessary.

[0056] Step S2: The gateway sequentially executes a sliding window on the received data. Outlier removal and exponentially weighted moving average filtering are performed. The processed data is then converted into a JSON format message, which is sent to the access point via Ethernet. The access point then forwards the message to the server via TCP / IP protocol.

[0057] Step S3: The server's stage identification module combines fermentation time, parameter range classification, parameter change trend, and anti-shake duration constraints to determine the current fermentation stage and matches the corresponding parameters from the stage parameter configuration table. , and The state assessment module calculates the S-value and stacking level of each node using the Gaussian membership function and weighted summation formula. The anomaly alarm module is for... The absolute value and rate of change are monitored, and alarm information is generated according to the three-level alarm rules when necessary.

[0058] Step S4: The server will store multi-dimensional parameters, stage information, Values ​​and alarm information are output to the display screen in a preset format, and visualized in the form of a 3D color temperature cloud map, parameter details panel, and historical curve panel. Alarm information is pushed in three levels: yellow warning, orange alarm, and red emergency alarm, respectively, via display pop-up window, SMS, and mobile APP. At the same time, all raw data, filtered data, Values ​​and alarm records are stored in the database in time series for historical tracing and fermentation process optimization analysis.

[0059] Example 2 The main difference between this embodiment and Embodiment 1 lies in the form of the normalized membership function used in the state evaluation module. In this embodiment, the normalized membership function... Represented using trigonometric functions, that is: when hour, ; when hour, ; when or hour, .

[0060] in, and The lower and upper limits of the parameter in the aforementioned stage parameter configuration table are taken respectively. This membership function form has a smaller computational load and is suitable for scenarios where server-side computing resources are relatively scarce. The remaining structure, deployment, preprocessing, stage identification, alarm rules, and visualization mapping are the same as in Example 1, and will not be repeated here.

[0061] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0062] The above embodiments merely illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A curved block fermentation status monitoring system, characterized by, Includes multiple sensor nodes, gateways, access points, and servers; Multiple sensor nodes are distributed and deployed at different locations in the curved block stack to collect multi-dimensional environmental parameters, including temperature, humidity, oxygen concentration and carbon dioxide concentration, and transmit the multi-dimensional environmental parameters to the gateway wirelessly. The gateway is used to perform protocol conversion and preprocessing on the received multidimensional environmental parameters, and then send the processed data to the access point. Access points are used to forward processed data to the server; The server is used to calculate the fermentation state index based on multidimensional environmental parameters, and outputs the multidimensional environmental parameters and fermentation state index to the display screen in a preset format for display.

2. The fermentation status monitoring system for koji blocks according to claim 1, characterized in that, Multiple sensor nodes are divided into several layers according to the vertical height direction of the curved block stack. Each layer includes at least two sensor nodes arranged at intervals along the horizontal direction. Each sensor node is equipped with a node identifier consisting of layer number, row number and column number. The node identifier corresponds one-to-one with the spatial position of the sensor node in the curved block stack. Multiple sensor nodes and the gateway form a star topology centered on the gateway, and the data frames between the multiple sensor nodes and the gateway all carry a CRC check field. The gateway sends a retransmission instruction for data frames that fail to be received, and the maximum number of retransmissions for a single frame does not exceed 3.

3. The fermentation status monitoring system for koji blocks according to claim 1, characterized in that, The sensor nodes and the gateway use the LoRa wireless communication protocol for data transmission. Data transmission between the gateway and the access point is performed using Ethernet or Wi-Fi communication protocols. Data transmission between the access point and the server is performed using the TCP / IP communication protocol. The sensor node requests SNTP time calibration from the server through the gateway every 24 hours after power-on and thereafter.

4. The fermentation status monitoring system for koji blocks according to claim 1, characterized in that, Gateway preprocessing includes 3D sliding window-based methods. Outlier removal and exponentially weighted moving average filtering; The length of the sliding window is , The value is a positive integer between 10 and 30. The sliding window is updated by rolling according to the first-in-first-out rule, and the mean of the sampled values ​​within the sliding window is... and standard deviation Calculate using the following formulas respectively: in Values ​​range from 0 to integers, These are sampled values; when At that time, Determined as an outlier and as described above It serves as a substitute value in subsequent filtering.

5. The fermentation status monitoring system for koji blocks according to claim 4, characterized in that, The initial value of the exponentially weighted moving average filter Take as the system power-on before The arithmetic mean of the next valid sample values It is a positive integer between 3 and 5.

6. The fermentation status monitoring system for koji blocks according to claim 1, characterized in that, The server includes a status assessment module, a phase identification module, and an anomaly alarm module. The state assessment module calculates the fermentation state index based on multidimensional environmental parameters according to the following formula: and Where the normalized membership function Using Gaussian functions: The stage identification module is used to divide the fermentation process of the koji blocks into several fermentation stages based on the changing trends of fermentation time and multidimensional environmental parameters, and dynamically adjusts the weighting coefficients according to the current fermentation stage. and optimal value ; The anomaly alarm module is used to detect an anomaly based on the fermentation state index. absolute value and rate of change per unit time Generate alarm information.

7. The fermentation status monitoring system for koji blocks according to claim 6, characterized in that, The stage identification module has a preset stage parameter configuration table. The stage parameter configuration table stores the fermentation time range, temperature upper and lower limits, humidity upper and lower limits, oxygen concentration upper and lower limits, carbon dioxide concentration upper and lower limits, and the corresponding optimal values ​​for each stage for each stage: mold growth, mold drying, damp fire, high fire, post-fire, and koji cultivation. Tolerance coefficient and weighting coefficients ; When the stage recognition module determines that the current environmental parameters meet the judgment conditions for a certain stage, the duration of the judgment conditions must also be no less than the preset anti-shake duration. Only then can the process switch to this stage. Take 15 to 60 minutes.

8. The fermentation status monitoring system for koji blocks according to claim 6, characterized in that, The rate of change per unit time In Take 15 to 60 minutes; The abnormal alarm module is based on the fermentation status index. The alarm levels are divided into three levels based on the size and rate of change, from light to severe: yellow alert, orange alert, and red emergency alert. Alarm information of different levels is pushed through display pop-ups, SMS platforms, and pre-bound terminal devices, respectively.

9. A method for monitoring the fermentation status of koji blocks, applied to the koji block fermentation status monitoring system as described in any one of claims 1 to 8, characterized in that, Includes the following steps: S1. Multiple sensor nodes collect multi-dimensional environmental parameters at different locations of the stacked curved blocks according to a preset sampling period, and send the multi-dimensional environmental parameters to the gateway wirelessly. S2. The gateway performs protocol conversion and preprocessing on the received multidimensional environmental parameters, and forwards the processed data to the server sequentially through the access point. S3. The server identifies the current fermentation stage, matches the corresponding weight coefficient and optimal value, calculates the fermentation state index based on multidimensional environmental parameters, and outputs the multidimensional environmental parameters and fermentation state index to the display screen in a preset format for display.

10. The method for monitoring the fermentation state of koji blocks according to claim 9, characterized in that, Step S3 is followed by the fermentation state index. Below the preset threshold or fermentation state index rate of change per unit time When the preset change rate threshold is exceeded, the server generates alarm information according to the alarm level and pushes it in a tiered manner through the display pop-up, SMS and pre-bound terminal devices respectively; at the same time, the server maps the multi-dimensional environmental parameters and fermentation state index to the stacking space of the briquette blocks shown on the display screen in the form of a three-dimensional color temperature cloud map according to the node identifier.

Images