Organic solid waste treatment line real-time monitoring system based on internet of things

Abstract

The application discloses an organic solid waste treatment line real-time monitoring system based on Internet of Things, which realizes real-time online sensing of apparent viscosity of non-Newtonian fluid by fusing multi-source sensor data by means of Internet of Things technology and realizing soft measurement modeling, on the basis of which, in combination with historical working condition envelope surface, prospective load risk assessment is carried out on the current dynamic, finally, through executing feedforward adaptive frequency control, malignant deviation tending to equipment failure is accurately identified and avoided, and the risk assessment lag and isotropic error existing in the traditional monitoring scheme are effectively solved. In this way, not only the false alarm and missed alarm rates can be significantly reduced, the transmission shaft torsional deformation and fracture can be effectively avoided, the operation safety of the large-scale stirring equipment can be ensured, but also the production interruption caused by frequent tripping is minimized, and the operation efficiency, the operation stability and the intelligent monitoring level of the organic solid waste treatment line are fundamentally improved.

Description

Technical Field

[0001] This application relates to the field of intelligent monitoring, and more specifically, to a real-time monitoring system for an organic solid waste treatment line based on the Internet of Things. Background Technology

[0002] With the increasing global demands for environmental governance and resource recycling, the efficient treatment and resource utilization of organic solid waste has become an important issue in the field of environmental engineering. In complex physicochemical processes, establishing a comprehensive real-time monitoring system for organic solid waste treatment lines is crucial for ensuring the safe operation of large-scale mixing equipment, improving material handling efficiency, and reducing economic losses caused by unexpected downtime.

[0003] Currently, most existing monitoring solutions for organic solid waste treatment lines rely on PLC systems to monitor basic electrical parameters such as motor current and voltage. However, because organic solid waste typically exhibits non-Newtonian fluid characteristics, its viscosity and rheology fluctuate drastically during treatment, leading to highly unstable motor loads on the treatment line. Existing technologies usually employ fixed protection current thresholds, which often result in delayed responses when the material becomes viscous, causing mechanical stress to exceed fatigue limits before the current triggers protection, leading to torsional deformation or even breakage of the drive shaft. Conversely, during normal fluctuations, frequent tripping and shutdowns can disrupt production continuity.

[0004] Therefore, an optimized IoT-based real-time monitoring system for organic solid waste treatment lines is needed. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides a real-time monitoring system for organic solid waste treatment lines based on the Internet of Things.

[0006] According to one aspect of this application, a real-time monitoring system for an organic solid waste treatment line based on the Internet of Things is provided, comprising: The data acquisition module is used to acquire real-time feedback data from the frequency converter, power of the mixing motor, and data from the feed displacement sensor. The synchronization feature vector construction module is used to construct synchronization feature vectors based on real-time feedback data from the frequency converter, power of the stirring motor, and data from the feed displacement sensor. The apparent viscosity soft measurement modeling module is used to perform apparent viscosity soft measurement modeling on synchronous feature vectors to obtain estimated values ​​of material apparent viscosity. The load risk index assessment module is used to assess the load risk index of the material's apparent viscosity estimate based on the historical normal operating condition envelope to obtain the load risk index. The feedforward adaptive frequency control module is used to perform feedforward adaptive frequency control on the load risk index to obtain an optimized feeding frequency.

[0007] Compared with existing technologies, this application provides an IoT-based real-time monitoring system for organic solid waste treatment lines. This system integrates multi-source sensor data using IoT technology and achieves real-time online sensing of the apparent viscosity of non-Newtonian fluids through soft-sensor modeling. Based on this, it performs a forward-looking load risk assessment of the current dynamics by combining historical operating condition envelopes. Finally, by executing feedforward adaptive frequency control, it accurately identifies and avoids malicious deviations leading to equipment failure, effectively solving the problems of risk assessment lag and isotropic fallacy inherent in traditional monitoring schemes. This approach not only significantly reduces false alarms and false negatives, effectively preventing drive shaft torsional deformation and breakage and ensuring the safe operation of large mixing equipment, but also minimizes production interruptions caused by frequent trips, fundamentally improving the operational efficiency, stability, and intelligent monitoring level of organic solid waste treatment lines. Attached Figure Description

[0008] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0009] Figure 1 This is a block diagram of an IoT-based real-time monitoring system for an organic solid waste treatment line according to an embodiment of this application. Figure 2 This is a schematic diagram of data flow in an IoT-based real-time monitoring system for an organic solid waste treatment line according to an embodiment of this application. Detailed Implementation

[0010] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0011] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0012] While this application makes various references to certain modules of the systems according to embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0013] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0014] The technical solution of this application proposes a real-time monitoring system for an organic solid waste treatment line based on the Internet of Things. Figure 1 This is a block diagram of an IoT-based real-time monitoring system for an organic solid waste treatment line according to an embodiment of this application. Figure 2 This is a schematic diagram of data flow in an IoT-based real-time monitoring system for an organic solid waste treatment line according to an embodiment of this application. Figure 1 and Figure 2 As shown, the IoT-based real-time monitoring system 300 for an organic solid waste treatment line according to an embodiment of this application includes: a data acquisition module 310, used to acquire real-time feedback data from the frequency converter, power of the stirring motor, and data from the feed displacement sensor; a synchronization feature vector construction module 320, used to construct a synchronization feature vector based on the real-time feedback data from the frequency converter, power of the stirring motor, and data from the feed displacement sensor; an apparent viscosity soft measurement modeling module 330, used to perform apparent viscosity soft measurement modeling on the synchronization feature vector to obtain an estimated value of the material's apparent viscosity; a load risk index assessment module 340, used to assess the load risk index on the estimated value of the material's apparent viscosity based on the historical normal operating condition envelope to obtain a load risk index; and a feedforward adaptive frequency control module 350, used to perform feedforward adaptive frequency control on the load risk index to obtain an optimized feeding frequency.

[0015] Specifically, the data acquisition module 310 is used to acquire real-time feedback data from the frequency converter, power of the stirring motor, and data from the feed displacement sensor. Since organic solid waste is a complex non-Newtonian fluid, its viscosity and rheology fluctuate drastically with operating conditions during processing, making direct measurement by conventional sensors impossible. Furthermore, the motor load is affected by multiple variables such as rotational speed, feed rate, and mechanical friction. Therefore, in this application's technical solution, soft measurement technology is employed to acquire various easily measurable variables that indirectly reflect the material state and equipment load, providing raw data input for subsequent viscosity estimation, risk assessment, and adaptive control. Specifically, the real-time feedback data from the frequency converter refers to the dynamic parameters such as operating frequency and speed transmitted back by the frequency converter. These parameters directly affect the determination of the stirring shear rate and provide a speed benchmark for the system to calculate material resistance. The stirring motor power refers to the total power consumed by the motor to overcome frictional resistance and fluid resistance under the current operating conditions. It is the core input for soft measurement modeling and provides an energy index for sensing changes in material viscosity. The feed displacement sensor data characterizes the feed rate in real time by measuring the stroke of the feeding device, enabling the system to identify anisotropic deviations such as "high viscosity and high feed," thereby avoiding the risk of drive shaft breakage due to a surge in load.

[0016] In practical implementation, the first step is to establish communication connections and configure data interfaces. The system establishes a communication connection with the frequency converter through an IoT gateway or industrial bus (such as Profinet, Modbus TCP / IP), and obtains real-time feedback data from the frequency converter through periodic polling or subscription. This data typically includes angular velocity, output torque, output current, and DC bus voltage, which characterize the rotational speed of the agitator blades. Simultaneously, the system needs to acquire the raw three-phase active power signal of the agitator motor during operation through an analog input module or power transmitter. This signal undergoes A / D conversion and scaling transformation to obtain the agitator motor power. Furthermore, the system acquires pulse signals or analog voltage / current signals emitted by the displacement sensor of the feeding mechanism (such as a screw feeder) through a digital input module or a dedicated displacement sensor reading head. By calculating the calibration relationship between the number of pulses or signal magnitude and the feed rate, it converts this data into feed displacement data, which directly reflects the feed rate per unit time.

[0017] Secondly, timestamp synchronization and data buffering are performed. Since the three types of data mentioned above may originate from different acquisition subsystems and suffer from communication delays and inconsistent sampling periods, each successfully acquired data point is assigned a high-precision timestamp (e.g., using the IEEE 1588 precision clock protocol to synchronize the clocks of each node), and these timestamps are stored in their respective temporary data buffers. This step forms the basis for subsequent time base unification and data alignment, ensuring the consistency of data points for different physical quantities across the time dimension.

[0018] Specifically, the synchronous feature vector construction module 320 is used to construct a synchronous feature vector based on real-time feedback data from the frequency converter, the power of the stirring motor, and the data from the feed displacement sensor. Since the sampling frequencies of the frequency converter feedback, motor power acquisition, and displacement sensor are often inconsistent, and the motor current is highly coupled with the four variables of feed rate, stirring speed, bearing friction, and material viscosity, the system cannot directly extract pure material resistance changes from the original electrical signal. Therefore, in the technical solution of this application, a synchronous feature vector is constructed to transform dispersed physical quantities into interrelated feature vectors in the same time dimension, providing high-fidelity, strongly correlated input features for subsequent soft-measurement inversion of apparent viscosity and feedforward control of load risk.

[0019] In practice, the following steps are taken: First, the real-time feedback data from the frequency converter, the power of the stirring motor, and the data from the feed displacement sensor are unified in time and resampled to obtain an aligned dataset. This aligned dataset includes frequency converter data, stirring power, and displacement data under a unified timestamp. Since the frequency converter, power transmitter, and displacement sensor may have different data refresh rates and communication delays, the system needs to unify them to the same time base. Specifically, the system uses a unified clock source (such as the PLC's system clock) as the reference and adds a high-precision timestamp to each newly arriving data point. Then, the system sets a master sampling period, using this period as the time grid, and resamples all data channels. For channels where no native data arrives at a certain grid point, linear interpolation is used to calculate an interpolation value based on the nearest data points before and after that channel, thus filling the data at that grid point. Through this operation, an aligned dataset is finally generated, containing a series of frequency converter data, stirring power, and displacement data under a unified timestamp, ensuring that the values ​​of all variables at any given time strictly correspond to the operating conditions at the same physical moment.

[0020] Secondly, a sliding window denoising process is applied to the original aligned sequence of stirring power in the aligned dataset to obtain a smooth power sequence. It should be understood that the stirring motor power signal, due to its electrical characteristics, is easily affected by instantaneous fluctuations and noise interference; direct use of this signal will affect the stability of subsequent models. Therefore, the original aligned sequence of stirring power in the aligned dataset is filtered. In this process, a filter with a length of... A sliding time window is used to perform neighborhood averaging denoising on the original power sequence. For each point in the sequence, the values ​​before and after it are taken. points ( The arithmetic mean of the values ​​of the power signal is calculated and used as the smoothed value for that point. This operation effectively suppresses high-frequency noise in the power signal, resulting in a smoothed power sequence that better reflects the true trend of power changes.

[0021] Furthermore, physical feature mapping and synchronization fusion vector construction are performed on the smoothed power sequence and the aligned inverter data sequence and displacement data sequence in the aligned dataset to obtain the synchronization feature vector. This step aims to fuse aligned and denoised time-series data with different physical units into a feature vector with comprehensive physical meaning. Specifically, the smoothed power sequence and the aligned displacement data sequence are used as inputs. At each unified timestamp, these scalar data are directly combined into a column vector, namely the synchronization feature vector, which contains three core physical quantities: total power, angular velocity, and feed rate.

[0022] Specifically, the apparent viscosity soft-sensing modeling module 330 is used to perform soft-sensing modeling of the apparent viscosity of the synchronous feature vector to obtain an estimated value of the material's apparent viscosity. It should be understood that organic solid waste is a typical non-Newtonian fluid, and its viscosity dynamically changes with water content, degradation degree, and stirring shear rate. Furthermore, the harsh industrial environment makes real-time online measurement impossible using traditional physical viscometers. In addition, the total power of the stirring motor is deeply coupled with mechanical factors such as bearing friction and transmission losses, and material resistance, resulting in the original electrical signal not directly reflecting the true changes in material properties. Through soft-sensing modeling, the system can decouple and inversely deduce the apparent viscosity of the material from observable electrical and kinematic parameters, providing crucial physical basis for subsequent load risk assessment and feed frequency control.

[0023] In practice, firstly, based on a preset loss characteristic function, the total power and angular velocity in the synchronization characteristic vector are calibrated under no-load conditions and dynamically stripped to obtain the net fluid breaking power. It should be understood that the total power consumed by the stirring motor is not only used to overcome the viscous resistance of the material, but also includes mechanical no-load losses such as bearing friction, gearbox wear, and wind resistance. To accurately obtain the net power used for breaking the material, the no-load loss is stripped from the total power. Specifically, no-load calibration and dynamic stripping are performed based on a preset loss characteristic function calibrated under no-load conditions. This function describes the no-load loss power of the system at a specific rotational speed, typically modeled as a quadratic polynomial related to angular velocity. In practice, the system extracts the total power and angular velocity at the current moment from the synchronization characteristic vector, calculates the no-load loss power at that rotational speed based on the current angular velocity using the loss characteristic function, and then subtracts the no-load loss power from the total power to obtain the net power used for fluid breaking. This process is expressed by the formula: ;

[0024] in, For total power, Angular velocity, , and For preset weighting coefficients, This represents the no-load power loss at the current speed. This refers to the net fluid crushing power.

[0025] Secondly, based on the system's geometric parameter set, the apparent viscosity is inverted from the net fluid breaking power and angular velocity to obtain an estimated value of the material's apparent viscosity. That is, after obtaining the net fluid breaking power, a mathematical relationship is established between it and the material's apparent viscosity based on the principles of agitated fluid dynamics. Specifically, for Newtonian fluids or materials that can be considered Newtonian fluids within a certain shear rate range, under turbulent conditions, the agitation power is proportional to the viscosity, the square of the rotational speed, and the cube of the agitator blade diameter. Based on this physical relationship, the system utilizes the system's geometric parameter set (mainly the characteristic diameter of the agitator blades) to... and geometric constants related to blade shape ), net fluid crushing power and angular velocity Inversion calculations are performed. Specifically, the apparent viscosity is inverted based on the net fluid breakup power and angular velocity using the following formula: ; in, Net fluid crushing power, Angular velocity, These are the geometric constants of the blade. This is the characteristic diameter of the agitator blade.

[0026] Specifically, the load risk index assessment module 340 is used to assess the load risk index of the estimated material apparent viscosity based on the historical normal operating condition envelope to obtain the load risk index. It should be understood that the existing mechanism has technical flaws in assessing the load risk index of the estimated material apparent viscosity based on the historical normal operating condition envelope. These flaws manifest in two aspects: First, the mechanism suffers from the isotropic fallacy in risk assessment. Mahalanobis distance measures the degree of anomaly by calculating the statistical distance from a point to the center of the data distribution, implicitly assuming that equidistant deviations in any direction represent the same level of risk. However, in the physical system of organic solid waste treatment, risk is highly anisotropic. Specifically, for the state vector composed of the material's apparent viscosity and feed rate, the direction of deviation has drastically different physical meanings. When the state point deviates towards the quadrant of low viscosity and low feed rate, it physically only means that the equipment is operating under low load or no load, which is a production efficiency issue, not a equipment safety risk. However, because it also deviates from the center of normal operating conditions, the Mahalanobis distance will give a large risk value, thus generating low-risk, high-false-alarm interference signals. Conversely, when the state point moves towards the quadrant of high viscosity and high feed rate, even if it has not reached the alarm threshold in terms of statistical distance, it may have already caused the torque of the stirring motor to rise sharply, approaching the critical point of stall or drive shaft distortion. Due to its directional blindness, the existing mechanism cannot identify this malicious deviation pointing towards the equipment failure area, resulting in the fatal flaw of high risk and low false alarm, misjudging potential catastrophic failures as fluctuations within the tolerance range. Secondly, the mechanism ignores the nonlinear coupling effect of key physical quantities. In the solid waste treatment process, the contribution of viscosity and feed rate to the system load is not a simple linear superposition, but a strong multiplicative effect. When the material viscosity is low, the increase in feed rate has a relatively mild impact on motor power; however, when the material viscosity rises to a certain threshold, even a small increase in feed rate will cause the system impedance to increase exponentially, rapidly pushing the system to an overloaded high-impedance region or a quasi-singularity point. Traditional Mahalanobis distance relies on the covariance matrix to describe the correlation between variables, which is essentially a linear representation and cannot capture the nonlinear coupling strength that increases dramatically with state changes. Therefore, existing mechanisms cannot predict that a system is approaching its physical limits along a steep risk surface; their assessments lag behind the actual physical process and lose their value as a warning system.

[0027] To address the aforementioned issues, this improved mechanism proposes a dynamic risk assessment method that integrates momentum correction and anisotropic projection penalty. By introducing the rate of state change and a directional penalty pointing towards the physical collapse zone, it constructs a comprehensive risk index that better reflects the actual physical processes in industrial settings.

[0028] In practice, firstly, based on the historical state cache queue, feature augmentation is performed on the estimated material apparent viscosity and the current feed rate in the synchronous feature vector to obtain an augmented dynamic state vector. This step aims to overcome the limitation of traditional static snapshots in capturing the direction of system evolution. Specifically, by obtaining the material apparent viscosity and feed rate at the current and previous sampling times, a velocity vector in the state space is calculated. This vector represents the instantaneous rate and direction of material properties and operating condition changes. Subsequently, this velocity vector is multiplied by a preset momentum prediction factor and added to the current state vector to generate a momentum-corrected augmented dynamic state vector. This augmented vector no longer merely describes the current position of the system but predicts the future position the system will reach under the current inertia, thus transforming risk assessment from a reactive response to a proactive prediction. This process is expressed by the formula:

[0029]

[0030] in, The state velocity vector represents the rate of change of the material's apparent viscosity and feed rate per unit time, reflecting the dynamic trend of the system's state evolution. This is the estimated value of the material's apparent viscosity at the current moment. This represents the current feed rate or related characteristic value. This represents the time interval between two samples. This represents the transpose operation of a vector. To augment the dynamic state vector, a predictive state representation that integrates the current state and future trends is provided. The momentum predictor is a time constant used to adjust the weight of the velocity component in predicting future states.

[0031] Next, based on the historical normal operating condition envelope, the mean vector and covariance matrix of the normal operating conditions are determined. Since organic solid waste treatment involves complex non-Newtonian fluid dynamics processes, its motor load is affected by the dual nonlinearity of material viscosity and feed rate. By determining the mean vector and covariance matrix of the normal operating conditions, the statistical distribution characteristics of normal production fluctuations can be quantified. This not only provides a mathematical basis for subsequent calculation of Mahalanobis distance to measure the degree of state outliers, but more importantly, it provides a reference center for identifying malignant deviations pointing to the high viscosity and high feed rate quadrants. This effectively distinguishes between harmless process fluctuations and critical risk trends that predict equipment failure (such as drive shaft torsion or stall).

[0032] In practice, the system first extracts historical normal operating condition envelope data from the storage unit. This envelope is a feature set consisting of estimated material apparent viscosity values, manually verified as safe, and the corresponding feed rate recorded during long-term stable operation. Next, the system trains using historical data to calculate the arithmetic mean of these multi-dimensional features, thereby determining the normal operating condition mean vector, which represents the center of the safe operating condition distribution. Subsequently, the system calculates the covariance between each dimension of the features, generating a covariance matrix. This matrix describes the variance and covariance between each dimension of the state vector, where variance measures the dispersion of a single variable, and covariance measures the linear correlation between two variables.

[0033] Furthermore, an anisotropic projection penalty is calculated on the mean vector of normal operating conditions and the augmented dynamic state vector, pointing towards the collapse zone, to obtain the anisotropic penalty factor. That is, to address the asymmetry of risk, an anisotropic projection penalty mechanism specifically targeting the physical hazard direction is introduced. It should be understood that risk is not uniformly distributed in the state space, and deviations towards the equipment failure area must be amplified. Specifically, firstly, a unit basis vector pointing towards the high-risk quadrant of high viscosity and high feed rate is predefined in the feature space, i.e., the blockage / collapse direction basis vector. Then, the deviation vector of the augmented state vector from the mean center of normal operating conditions is calculated, and this deviation vector is projected onto the preset hazard direction basis vector. To ensure that only dangerous trends are penalized, a linear rectified function (ReLU) is used to process the projected value, so that deviations towards the safe region do not introduce penalty. Subsequently, this non-negative projected value is nonlinearly amplified through an exponential function to simulate the avalanche effect when the system approaches the physical critical point, thereby obtaining the anisotropic penalty factor. This imposes physical constraints on the purely statistical model, constructing a risk potential field whose intensity increases sharply in the dangerous direction, thus precisely solving the problem that the original mechanism cannot identify the fatal direction. This process can be expressed by the following formula:

[0034]

[0035] in, The projection component of the deviation vector in the dangerous direction quantifies the tendency of the current state to move toward the congested area. This is the mean vector of normal operating conditions obtained through training with historical data. The preset, normalized blockage / collapse direction basis vector represents the gradient direction with the highest risk of system overload failure. For anisotropy penalty factor, a dimensionless risk amplification coefficient. This is a non-linear sensitivity coefficient used to adjust the severity of risk penalties. It is a linear rectified function, when the input... Output 0 when Time output .

[0036] Subsequently, a comprehensive risk index is synthesized from the anisotropy penalty factor, the augmented dynamic state vector, the inverse of the covariance matrix, and the mean vector of the normal operating condition to obtain the load risk index. That is, a universal statistical distance is fused with a targeted physical field penalty to generate a comprehensive and dynamic comprehensive risk index, thus taking into account both the statistical understanding of the overall operating condition distribution and the precise amplification of specific physical risks. Specifically, firstly, the Mahalanobis distance of the augmented state vector relative to the historical normal operating condition distribution is calculated, which effectively measures the statistically significant outlier degree of the current state. Then, this Mahalanobis distance is multiplied by the anisotropy penalty factor to obtain the final anisotropic dynamic load risk index. Thus, through multiplicative coupling, the risk index only significantly increases when the state point simultaneously satisfies both statistical anomalies and physical movement towards a dangerous direction, thereby greatly improving the accuracy and reliability of the early warning. This process is expressed by the formula:

[0037]

[0038] in, It provides a basic statistical measure of the degree of anomaly for Mahalanobis distance calculated based on augmented vectors. It is the inverse of the covariance matrix of the normal operating condition data. The final output is the anisotropic dynamic load risk index, a risk metric that combines statistical outlier and physical hazard direction.

[0039] In summary, this improved mechanism generates a dynamic index that accurately reflects the true risk level of the organic solid waste treatment line's load. Specifically, by incorporating momentum from state changes and directional penalties pointing towards physical failure zones, this risk index effectively distinguishes between harmless data deviations caused by normal process fluctuations and critical risk trends that truly foreshadow equipment stalling or overload. Compared to existing mechanisms, this improved mechanism significantly reduces false alarm and false negative rates, enabling the system to identify potential equipment safety threats earlier and more accurately. Ultimately, it provides a reliable decision-making basis for implementing preventative adaptive control (such as automatically reducing the feed rate when the risk index increases), thereby fundamentally improving the operational safety, stability, and overall treatment efficiency of the organic solid waste treatment production line and effectively avoiding significant economic losses caused by unexpected equipment downtime or damage.

[0040] Specifically, the feedforward adaptive frequency control module 350 is used to perform feedforward adaptive frequency control on the load risk index to obtain an optimized feeding frequency. It should be understood that although the load risk index quantitatively reflects the current or predicted system load level, without corresponding adjustments to the feeding process, a high-risk index can only serve as a warning and cannot prevent potential equipment stalling or damage. Therefore, in the technical solution of this application, feedforward adaptive frequency control is adopted to adjust the control quantity (feeding frequency) in advance based on the real-time assessed risk (i.e., the estimated effect of interference), mapping the abstract load risk index to a specific, executable optimized feeding frequency. This automatically reduces the feed rate when the risk increases and restores or increases the feed rate when the risk decreases, ultimately achieving a dynamic balance between ensuring equipment safety and optimizing production efficiency.

[0041] In practice, firstly, based on a preset risk alarm threshold, the load risk index is assessed for load risk threshold determination and over-limit measurement to obtain the risk over-limit value. Specifically, a preset risk alarm threshold represents the critical risk level at which the system enters a state requiring intervention. When the calculated load risk index is below this threshold, it indicates that the system is in a safe state and no adjustment of the feeding frequency is needed; in this case, the risk over-limit value should be set to zero. Only when the load risk index exceeds this threshold is an overload risk considered to exist, requiring the activation of the control algorithm. The excess portion is used as the controller input, and its magnitude determines the amplitude of frequency adjustment. This process converts the continuous risk index into a non-negative over-limit quantity, providing a clear, unidirectional driving signal for subsequent control law calculations.

[0042] Secondly, an exponential decay control law is applied to the risk exceedance value, the basic feeding frequency, and the sensitivity coefficient to obtain the optimized feeding frequency. That is, an exponential decay control law is used to generate the final feeding frequency command. The input parameters of this control law include the risk exceedance value calculated in the previous step, a basic feeding frequency set by the process (representing the target feeding frequency the system expects to achieve when there is no risk or the risk is negligible), and a control sensitivity coefficient. Specifically, the feeding frequency should decrease exponentially with the increase of the risk exceedance value, based on the basic frequency. This exponential relationship enables a non-linear response to risk, i.e., mild suppression when the risk just exceeds the limit, and strong intervention when the risk rises sharply, meeting the stringent safety requirements of industrial processes. Specifically, the exponential decay control law is applied to the risk exceedance value, the basic feeding frequency, and the sensitivity coefficient using the following formula:

[0043] in, The risk exceeds the limit. Based on the basic feeding frequency, A sensitivity coefficient greater than zero (used to adjust the rate of frequency decay as risk increases). The optimized feed frequency is the final output. Based on the natural constant An exponential function with base 0.05. This means that when... When the output frequency equals the fundamental frequency, ;when At that time, the output frequency decreases exponentially, and The larger the value, the faster the decay.

[0044] Taking the scheme in this application as an example, the system presets a risk alarm threshold of 3.0, and the basic feeding frequency is set according to process requirements. The sensitivity coefficient was set based on on-site debugging experience. At a certain sampling time The load risk index calculated by the system is First, determine the risk threshold and measure any exceedances: Next, the exponential decay control law is solved: The system then sends this frequency command to the frequency converter of the feed motor. This means that, due to the current high load risk index (4.2), the system automatically adjusts the feeding frequency from the ideal... Significantly reduced to approximately This is done to preventively reduce system load, thereby ensuring equipment safety. If the risk index drops to 2.5 (below the threshold) in the next moment, then... The feeding frequency will immediately return to 35 Hz.

[0045] As described above, the IoT-based real-time monitoring system 300 for organic solid waste treatment lines according to embodiments of this application can be implemented in various wireless terminals, such as servers with IoT-based real-time monitoring algorithms for organic solid waste treatment lines. In one possible implementation, the IoT-based real-time monitoring system 300 for organic solid waste treatment lines according to embodiments of this application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the IoT-based real-time monitoring system 300 can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the IoT-based real-time monitoring system 300 can also be one of many hardware modules of the wireless terminal.

[0046] Alternatively, in another example, the IoT-based real-time monitoring system 300 for organic solid waste treatment lines and the wireless terminal can also be separate devices, and the IoT-based real-time monitoring system 300 for organic solid waste treatment lines can be connected to the wireless terminal via wired and / or wireless networks, and transmit interactive information in accordance with an agreed data format.

[0047] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A real-time monitoring system for an organic solid waste treatment line based on the Internet of Things, characterized in that, include: The data acquisition module is used to acquire real-time feedback data from the frequency converter, power of the mixing motor, and data from the feed displacement sensor. The synchronization feature vector construction module is used to construct synchronization feature vectors based on real-time feedback data from the frequency converter, power of the stirring motor, and data from the feed displacement sensor. The apparent viscosity soft measurement modeling module is used to perform apparent viscosity soft measurement modeling on synchronous feature vectors to obtain estimated values ​​of material apparent viscosity. The load risk index assessment module is used to assess the load risk index of the material's apparent viscosity estimate based on the historical normal operating condition envelope to obtain the load risk index. The feedforward adaptive frequency control module is used to perform feedforward adaptive frequency control on the load risk index to obtain an optimized feeding frequency.

2. The real-time monitoring system for an organic solid waste treatment line based on the Internet of Things as described in claim 1, characterized in that, The synchronous feature vector construction module is used for: The real-time feedback data of the frequency converter, the power of the stirring motor and the data of the feed displacement sensor are unified and resampled to obtain an aligned dataset. The aligned dataset includes frequency converter data, stirring power and displacement data under a unified timestamp. Sliding window denoising is applied to the original aligned sequence of stirring power in the aligned dataset to obtain a smooth power sequence; Physical feature mapping and synchronization fusion vector construction are performed on the smoothed power sequence and the frequency converter data alignment sequence and displacement data alignment sequence in the aligned dataset to obtain the synchronization feature vector.

3. The real-time monitoring system for an organic solid waste treatment line based on the Internet of Things as described in claim 1, characterized in that, The apparent viscosity soft measurement modeling module includes: The no-load calibration and dynamic stripping unit is used to perform no-load calibration and dynamic stripping of the total power and angular velocity in the synchronous characteristic vector based on a preset loss characteristic function to obtain the net fluid breaking power. The apparent viscosity inversion unit is used to perform apparent viscosity inversion on net fluid crushing power and angular velocity based on the system geometric parameter set to obtain an estimated value of the material's apparent viscosity.

4. The real-time monitoring system for an organic solid waste treatment line based on the Internet of Things according to claim 3, characterized in that, The no-load calibration and dynamic stripping unit is used to perform no-load calibration and dynamic stripping of the total power and angular velocity in the synchronization characteristic vector using the following formula: ； ； in, For total power, Angular velocity, , and These are preset weighting coefficients.

5. The real-time monitoring system for an organic solid waste treatment line based on the Internet of Things according to claim 3, characterized in that, The apparent viscosity inversion unit is used to invert the apparent viscosity of the net fluid breakup power and angular velocity using the following formula: ； in, Net fluid crushing power, Angular velocity, These are the geometric constants of the blade. This is the characteristic diameter of the agitator blade.

6. The real-time monitoring system for an organic solid waste treatment line based on the Internet of Things according to claim 1, characterized in that, The load risk index assessment module is used for: Based on the historical state cache queue, feature augmentation is performed on the material apparent viscosity estimate and the current feed rate in the synchronous feature vector to obtain the augmented dynamic state vector. Based on the historical normal operating condition envelope, determine the mean vector and covariance matrix of the normal operating condition; Anisotropic projection penalty calculations are performed on the mean vector of normal working conditions and the augmented dynamic state vector, pointing towards the collapse zone, to obtain the anisotropic penalty factor. The load risk index is obtained by synthesizing a comprehensive risk index from the anisotropic penalty factor, the augmented dynamic state vector, the inverse of the covariance matrix, and the mean vector of normal operating conditions.

7. The real-time monitoring system for an organic solid waste treatment line based on the Internet of Things according to claim 1, characterized in that, The feedforward adaptive frequency control module includes: The load risk threshold determination and over-limit measurement unit is used to determine the load risk threshold and measure the over-limit based on the preset risk alarm threshold to obtain the risk over-limit value. The exponential decay control law calculation unit is used to calculate the exponential decay control law for the risk over-limit value, the basic feeding frequency and the sensitivity coefficient to obtain the optimized feeding frequency.

8. The real-time monitoring system for an organic solid waste treatment line based on the Internet of Things according to claim 7, characterized in that, The exponential decay control law calculation unit is used to calculate the exponential decay control law for the risk exceedance value, the basic feeding frequency, and the sensitivity coefficient using the following formula: ； in, The risk exceeds the limit. Based on the feeding frequency and This is the sensitivity coefficient.

Images