Multi-timing mode ultrasonic cleaning intelligent control system

Abstract

This invention discloses a multi-timing mode intelligent control system for ultrasonic cleaning, relating to the field of ultrasonic cleaning technology. It includes a cleaning condition modeling module, a multi-mode timing configuration module, a multi-source data fusion module, a resonant frequency adaptive tracking module, an abnormal state detection module, and an intelligent closed-loop control module. The cleaning condition modeling module acquires transducer vibration field data, constructs a three-dimensional sound field model using relevant algorithms, optimizes the transducer layout, and ensures the uniformity of the sound field within the cleaning tank. By setting up the cleaning condition modeling module, combined with a laser Doppler vibrometer, impedance sensor array, and PVDF sound intensity sensor, high-precision acquisition of transducer vibration field and sound intensity data is achieved. A three-dimensional sound field model is constructed based on finite element analysis, and a genetic algorithm is introduced to iteratively optimize the transducer installation angle and spacing. Combined with the temperature compensation circuit built into the sensor, automatic correction for ambient temperatures ranging from 0-80℃ is achieved.

Description

Technical Field

[0001] This invention relates to the field of ultrasonic cleaning technology, and in particular to an intelligent control system for ultrasonic cleaning with multiple timing modes. Background Technology

[0002] In modern industrial production, ultrasonic cleaning equipment serves as a core device for cleaning precision parts. Utilizing the high-frequency vibrations generated by the cavitation effect, it efficiently removes dirt, playing an indispensable role in ensuring the precision and quality of products in industries such as precision electronics, machinery manufacturing, medical devices, and automotive parts. Its working principle is based on the propagation of ultrasonic vibrations. A transducer converts electrical signals into mechanical vibrations, which are then transmitted to the cleaning fluid, forming dense cavitation bubbles. The impact force generated by the collapse of these cavitation bubbles cleans the surfaces of parts and complex structures such as blind holes and crevices. By adjusting parameters such as the transducer's drive frequency and operating time, it can adapt to the cleaning needs of different types of parts, making it a crucial piece of equipment for achieving precision cleaning in industrial production.

[0003] Currently, although various ultrasonic cleaning control systems exist on the market, significant technical bottlenecks remain. Some products lack precise modeling and optimization of the acoustic field distribution in the cleaning tank, have crude transducer layout designs, poor acoustic field uniformity, and incomplete cavitation effect coverage, resulting in poor cleaning effects and low cleaning consistency for precision parts. Others only support a single timed cleaning mode, with cumbersome parameter configuration and no support for remote control, slow mode switching response, and inability to adapt to the cleaning needs of different scenarios such as vulnerable parts and heavily soiled complex parts. Still other control systems lack a complete multi-source data acquisition and fusion mechanism, cannot track the transducer resonance state in real time, experience lag in drive frequency adjustment when the load changes, and have insufficient anomaly detection and closed-loop control capabilities, resulting in poor equipment stability and high energy consumption. Furthermore, inadequate designs in cleaning fluid management, electromagnetic protection, and energy-saving operation further limit the intelligence level and industrial adaptability of ultrasonic cleaning equipment. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent control system for ultrasonic cleaning with multiple timing modes, which solves the technical problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent control system for ultrasonic cleaning with multiple timing modes, comprising a cleaning condition modeling module, a multi-mode timing configuration module, a multi-source data fusion module, a resonant frequency adaptive tracking module, an abnormal state detection module, and an intelligent closed-loop control module; The cleaning condition modeling module is used to acquire transducer vibration field data, construct a three-dimensional sound field model in combination with relevant algorithms, optimize the transducer layout, ensure the uniformity of the sound field in the cleaning tank, provide a basic sound field model support for subsequent cleaning control, and output the optimized three-dimensional sound field model and transducer layout parameters. The multi-mode timing configuration module is used to preset multiple core timing modes, supports user-defined timing-related parameters, realizes real-time parsing, storage and retrieval of mode parameters, can complete rapid mode switching, adapt to different cleaning scenario requirements, and output parsed mode parameters and switching control signals. The multi-source data fusion module is used to collect multi-dimensional operational data during the cleaning process, perform noise reduction, registration and deviation correction on the collected data, eliminate data noise and deviation interference, integrate into standardized data, provide reliable data support for subsequent resonance tracking, anomaly detection and closed-loop control, and output clean fusion operation data. The resonant frequency adaptive tracking module is used to monitor the resonant state of the transducer in real time, dynamically adjust the driving frequency to compensate for load changes during the cleaning process, maintain the resonant stability of the transducer, ensure the acoustic energy transmission efficiency, and output the adjusted driving frequency parameters and resonant state monitoring data. The abnormal state detection module is used to construct a baseline feature space for cleaning parameters, and combined with a preset abnormal judgment threshold, identify abnormal situations during the operation of the cleaning equipment, promptly capture abnormal fluctuations in temperature, liquid level, and impedance parameters, and output abnormal detection results and alarm trigger signals. The intelligent closed-loop control module is used to link the start-stop control module when receiving an anomaly detection signal, dynamically adjust the system operating parameters, balance the cleaning effect and energy consumption, ensure that the equipment can quickly resume stable operation after the anomaly is handled, and output closed-loop control commands and optimized operating parameters.

[0006] Preferably, the cleaning condition modeling module includes a vibration field data acquisition unit, a three-dimensional sound field modeling unit, and a transducer layout optimization unit. The vibration field data acquisition unit acquires transducer vibration field data through a laser Doppler vibration meter and an impedance sensor array. Simultaneously, a PVDF sound intensity sensor is used to set sound intensity monitoring points throughout the cleaning tank. The sensor has a built-in temperature compensation circuit, which can automatically correct the influence of ambient temperature (0-80℃) on the measurement results, ensuring measurement accuracy ≤±2%. The temperature compensation circuit is calibrated experimentally to determine the compensation coefficient and outputs the original vibration field data and sound intensity monitoring data. The three-dimensional sound field modeling unit combines the finite element analysis method to analyze and process the collected vibration field data and sound intensity monitoring data, construct a three-dimensional sound field model, and use mesh adaptive technology to densify the edge mesh of the cleaning tank. The mesh densification criterion is to start when the edge sound field gradient exceeds the preset range, and output the three-dimensional sound field model. The transducer layout optimization unit introduces a genetic algorithm, sets reasonable fitness functions, constraints and iteration parameters, and iteratively optimizes the installation angle and spacing of the transducers to make the sound field uniformity error ≤8% and the cavitation effect coverage ≥95%. The sound field uniformity error and cavitation effect coverage are verified by corresponding test methods, and the optimized transducer layout parameters are output.

[0007] Preferably, the multi-mode timing configuration module includes a core mode preset unit, a parameter parsing and storage unit, and a mode switching unit; The core mode preset unit presets three core timing modes: continuous timing, pulse timing, and stepped timing. The continuous timing mode supports continuous operation from 10s to 180min, with the frequency adaptively adjusted following the resonant frequency and the reference frequency being 28kHz±2%. The pulse timing mode has a minimum pulse period of 100ms. The stepped timing mode supports 3-5 independent parameter configuration steps and outputs preset core timing mode parameters. The parameter parsing and storage unit uses an ARM Cortex-M4 processor to parse and store the timing mode parameters in real time. It supports parameter retention even when power is off. At the same time, it can remotely modify and batch configure parameters via RS485 bus. Remote communication adopts a standard communication protocol and a fixed data frame format, and outputs the parsed parameters and storage feedback signals. The mode switching unit adopts a fast response mechanism. The pulse trigger signal is generated by the processor timer module, and the step-by-step switching adopts a smooth transition algorithm. It can quickly switch between three core timing modes according to user instructions or cleaning scenario requirements, and output mode switching control signals.

[0008] Preferably, the multi-source data fusion module includes a multi-dimensional data acquisition unit, a data preprocessing unit, and a data output unit; The multi-dimensional data acquisition unit collects multi-dimensional operational data during the cleaning process, including cleaning fluid temperature, liquid level, transducer impedance, and sound intensity distribution data. The liquid level detection uses a magnetostrictive sensor, and the temperature detection uses a platinum resistance element. The sensors are set with a fixed sampling period and output multi-dimensional raw operational data. The data preprocessing unit uses a 5×5 median filtering algorithm to denoise the raw running data, completes data registration through an iterative nearest point algorithm, and corrects sound field deviations by combining a finite element sound field correction algorithm. The data fusion adopts a reasonable timing and weight allocation strategy to ensure that the parameter transmission delay is ≤50ms. The transmission delay is guaranteed by a caching optimization strategy, and the output is clean fused running data. The data output unit synchronously transmits the cleanroom integration operation data to the resonant frequency adaptive tracking module and the abnormal state detection module, providing data support for subsequent control links and outputting data transmission feedback signals.

[0009] Preferably, the resonant frequency adaptive tracking module includes a resonant state monitoring unit, a drive frequency adjustment unit, and a temperature compensation unit; The resonant state monitoring unit is based on a DSP processor and monitors the resonant state of the transducer in real time. It uses the impedance phase angle as the core monitoring index and outputs the transducer resonant state data. The drive frequency adjustment unit adopts the gradient approximation method, sets reasonable step size and convergence conditions, and dynamically adjusts the drive frequency to compensate for load changes based on the transducer resonance state data. Load compensation is completed within 0.3s, ensuring that the resonance point deviation is ≤±100Hz. The drive circuit adopts a full-bridge topology and soft-switching technology. The power MOSFET uses third-generation semiconductor materials, and the overall board efficiency is ≥92%. The soft-switching technology sets the corresponding dead time and drive strategy, and outputs the adjusted drive frequency parameters. The temperature compensation unit performs temperature coefficient compensation every 15 minutes, using a calibrated mathematical model to correct the influence of temperature changes on the resonance state, ensuring that the sound energy transmission efficiency is ≥90%, and outputs a temperature compensation signal.

[0010] Preferably, the abnormal state detection module includes a baseline feature space construction unit, an abnormal identification unit, and an abnormal reporting unit; The benchmark feature space construction unit is based on the K-means clustering algorithm, using sufficient normal working condition data as the training sample set. The number of clusters is determined by the elbow rule. A clean parameter benchmark feature space is constructed, and dimensionality reduction is performed by principal component analysis to retain ≥95% information entropy. The principal component selection is based on the eigenvalue setting, and the benchmark feature space parameters are output. The anomaly identification unit presets anomaly judgment thresholds, including temperature gradient > 3℃ / min, liquid level fluctuation > 5mm, and impedance change > 15%. These anomaly thresholds are calibrated through multiple experiments and combined with the benchmark feature space parameters to identify equipment malfunctions and output anomaly detection results. After identifying an anomaly, the anomaly reporting unit reports the anomaly event synchronously through local audible and visual alarms and remote communication. The alarm and reporting settings include fixed response delays and fault code rules, and the unit outputs an anomaly reporting signal.

[0011] Preferably, the intelligent closed-loop control module includes an optimization target setting unit, an operating parameter adjustment unit, and a stability recovery unit; The optimization target setting unit adopts a model predictive control algorithm, with the goal of "maximum cleaning efficiency and minimum energy consumption", establishes a target function with unified dimensions, sets reasonable weight coefficients for the target function, and outputs the optimization target parameters; The operating parameter adjustment unit adopts a rolling time-domain optimization strategy, sets the corresponding prediction step size and control cycle, and dynamically adjusts the system amplitude, frequency and running time. With the help of an electric regulating valve and a water pump with a response time of <50ms, the liquid level adjustment is completed within 200ms. The liquid level adjustment adopts a PID algorithm and outputs the adjusted operating parameters and control commands. After the anomaly is handled, the stability recovery unit ensures that the system returns to stable operation within 3 seconds. The stable operation is defined by clear judgment criteria and outputs a system stable operation feedback signal.

[0012] Preferably, it also includes a cleaning tank and a transducer structure module, wherein the cleaning tank and transducer structure module includes a cleaning tank structure unit and a transducer unit. The cleaning tank structural unit is made of ZL114A aluminum alloy in one piece and is prepared by die casting process. Microchannels are embedded in the inner wall to enhance the heat exchange effect. The microchannels are set with fixed cross-sectional size and spacing. The connection is a sealed joint with IP68 protection level, which is suitable for wide temperature conditions of -40℃ to 125℃. The cleaning tank structural parameters and protection performance parameters are output. The transducer unit uses lead scandate-lead titanate piezoelectric material. The composition ratio, preparation process and polarization conditions of the piezoelectric material are specified. The transducer has an electroacoustic conversion efficiency of ≥95% after continuous operation at 80℃ for 2000 hours. The electroacoustic conversion efficiency and protection performance are verified by corresponding standard tests, and the transducer performance parameters are output.

[0013] Preferably, it also includes a cleaning fluid management module, which includes a conductivity monitoring unit, a fluid replenishment and compensation unit, and a dead-angle-free cleaning unit; The conductivity monitoring unit uses a four-electrode conductivity sensor with a fixed measurement range and accuracy to monitor the conductivity of the cleaning fluid in real time. When the conductivity is >50μS / cm, the filtration device is automatically started. The filtration device is set with a fixed filtration cycle and a stop threshold, and outputs conductivity monitoring data and filtration control signals. The liquid replenishment unit has a built-in liquid replenishment device, which is linked to the liquid level sensor to realize automatic compensation of cleaning fluid. It can set the corresponding liquid replenishment threshold and speed, and output liquid replenishment control signal and liquid level monitoring data. The no-dead-angle cleaning unit uses a low-viscosity synthetic heat-conducting fluid, with clearly defined viscosity, thermal conductivity and applicable temperature range. It works in conjunction with circumferential emission transducers to achieve 360° no-dead-angle cleaning. The circumferential emission transducers are arranged in a fixed number and at a fixed angle to output cleaning control parameters.

[0014] Preferably, it also includes a protection and energy-saving operation module, which includes an EMC protection unit, a data storage unit, and an energy-saving control unit; The EMC protection unit adopts an industrial-grade EMC protection design, including power filtering, signal shielding and grounding treatment, to ensure the system's anti-interference capability and output EMC protection parameters. The data storage unit is equipped with an edge computing module with a processing latency of ≤100ms. It supports local storage of 1,000 sets of cleaning logs and parameters. The edge computing module uses an appropriate processor and storage medium to output data storage feedback signals. The energy-saving control unit sets clear energy-saving mode switching conditions and power reduction ratios, automatically switching to energy-saving mode during low-load periods, making the system suitable for cleaning in multiple scenarios, with a cleaning qualification rate of ≥97%, which is verified by experiments, and outputs energy-saving control commands and operation feedback signals.

[0015] Compared with related technologies, the ultrasonic cleaning intelligent control system with multiple timing modes provided by the present invention has the following beneficial effects: 1. This invention provides an intelligent control system for ultrasonic cleaning with multiple timing modes. By setting up a cleaning condition modeling module, combined with a laser Doppler vibrometer, impedance sensor array, and PVDF sound intensity sensor, it achieves high-precision acquisition of transducer vibration field and sound intensity data. Based on finite element analysis, a three-dimensional sound field model is constructed, and a genetic algorithm is introduced to iteratively optimize the transducer installation angle and spacing. With the temperature compensation circuit built into the sensor, automatic correction for ambient temperature of 0-80℃ is achieved. This solves the problems of uneven sound field distribution, large temperature influence on measurement accuracy, and incomplete cavitation effect coverage in traditional ultrasonic cleaning equipment, which lead to poor cleaning effect and low cleaning consistency. It achieves a sound field uniformity error of ≤8% in the cleaning tank, a cavitation effect coverage rate of ≥95%, and a measurement accuracy of ≤±2%, providing high-precision and high-reliability sound field basic model support for cleaning control.

[0016] 2. This invention provides an intelligent ultrasonic cleaning control system with multiple timing modes. By designing a multi-mode timing configuration module and building a full-link intelligent control system, it presets three core timing modes: continuous, pulse, and stepped. It supports parameter customization, power-off data saving, and remote batch configuration, and achieves seamless mode switching with a rapid response mechanism. At the same time, it standardizes the multi-dimensional cleaning data to achieve dynamic tracking and accurate compensation of the transducer resonant frequency. Combined with the algorithm to construct a reference feature space, it completes accurate identification of equipment anomalies and dual-end alarms. This solves the problems of traditional cleaning equipment having single modes, poor adaptability, cumbersome parameter configuration, and lack of real-time monitoring, dynamic tuning, and anomaly control during the cleaning process, resulting in low equipment stability and high manual maintenance costs. It achieves intelligent and precise control of the cleaning process.

[0017] 3. This invention provides an intelligent ultrasonic cleaning control system with multiple timing modes. By adding structural, media management, and energy-saving protection modules, it achieves a comprehensive upgrade of the equipment, optimizes the materials and structure of the cleaning tank and transducer, and improves the equipment's durability and wide temperature adaptability. It enables the monitoring of the conductivity of the cleaning fluid, automatic fluid replenishment, and 360° cleaning without dead angles. Combined with industrial-grade EMC protection, edge data storage, and intelligent energy-saving modes, it solves the problems of poor structural durability, rough management of cleaning fluid that easily leaves dead angles, weak anti-interference ability, high energy consumption, and lack of data traceability in traditional cleaning equipment. It achieves a cleaning qualification rate of ≥97% and reduces energy consumption by 45%-50% in energy-saving mode, significantly improving the equipment's environmental adaptability, energy efficiency, and traceability, and adapting to the cleaning needs of multiple industries. Attached Figure Description

[0018] Figure 1 This is a flowchart of the present invention; Figure 2 This is an extended flowchart of the cleaning condition modeling module of the present invention; Figure 3 This is an extended flowchart of the multi-mode timing configuration module of the present invention; Figure 4 This is an extended flowchart of the multi-source data fusion module of the present invention; Figure 5 This is an extended flowchart of the resonant frequency adaptive tracking module of the present invention; Figure 6 This is an extended flowchart of the abnormal state detection module of the present invention; Figure 7 This is an extended flowchart of the intelligent closed-loop control module of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] Example 1: Please see Figures 1-7 The present invention provides a technical solution: an ultrasonic cleaning intelligent control system with multiple timing modes, including a cleaning condition modeling module, a multi-mode timing configuration module, a multi-source data fusion module, a resonant frequency adaptive tracking module, an abnormal state detection module, and an intelligent closed-loop control module. The cleaning condition modeling module is used to acquire transducer vibration field data, construct a three-dimensional sound field model in combination with relevant algorithms, optimize the transducer layout, ensure the uniformity of the sound field in the cleaning tank, provide a basic sound field model support for subsequent cleaning control, and output the optimized three-dimensional sound field model and transducer layout parameters. The cleaning condition modeling module includes a vibration field data acquisition unit, a three-dimensional sound field modeling unit, and a transducer layout optimization unit. The vibration field data acquisition unit acquires transducer vibration field data through a laser Doppler vibration meter and an impedance sensor array. Simultaneously, a PVDF sound intensity sensor is used to set sound intensity monitoring points throughout the cleaning tank. The sensor has a built-in temperature compensation circuit, which can automatically correct the influence of ambient temperature (0-80℃) on the measurement results, ensuring measurement accuracy ≤±2%. The temperature compensation circuit is calibrated experimentally to determine the compensation coefficient and outputs the original vibration field data and sound intensity monitoring data. The three-dimensional sound field modeling unit combines the finite element analysis method to analyze and process the collected vibration field data and sound intensity monitoring data, construct a three-dimensional sound field model, and use mesh adaptive technology to densify the edge mesh of the cleaning tank. The mesh densification criterion is to start when the edge sound field gradient exceeds the preset range, and output the three-dimensional sound field model. The transducer layout optimization unit introduces a genetic algorithm, sets reasonable fitness functions, constraints and iteration parameters, and iteratively optimizes the installation angle and spacing of the transducers to make the sound field uniformity error ≤8% and the cavitation effect coverage ≥95%. The sound field uniformity error and cavitation effect coverage are verified by corresponding test methods, and the optimized transducer layout parameters are output. In this implementation scheme, the vibration field data acquisition unit synchronously acquires vibration displacement detection data from the laser Doppler vibrometer and impedance characteristic data from the impedance sensor array. The sampling frequency is consistent with that of the PVDF sound intensity sensor to ensure time dimension matching of multi-source detection data. The compensation coefficient of the temperature compensation circuit is obtained experimentally through calibration at 5℃ nodes within the 0-80℃ range, and accurate compensation is achieved across the entire temperature range through interpolation algorithms. The finite element analysis of the three-dimensional sound field modeling unit is based on the COMSOL multiphysics simulation platform. The mesh adaptive technology sets the sound field gradient threshold to 5% / mm. When the sound field gradient in the edge region exceeds this threshold, the mesh accuracy is automatically increased to four times the original accuracy to ensure the accuracy of edge sound field modeling. In the genetic algorithm of the transducer layout optimization unit, the fitness function uses sound field uniformity error and cavitation effect coverage as the core evaluation indicators. The number of iterations is set to 200 generations, the crossover probability is 0.8, and the mutation probability is 0.05. The transducer installation angle range is constrained to 0-90° and the spacing range is 50-200mm. After iterative optimization, the sound field uniformity error in the cleaning tank is stably controlled at 6%-8%, and the cavitation effect coverage can reach 95%-98%, providing a high-precision sound field basic model for subsequent cleaning control.

[0021] The multi-mode timing configuration module is used to preset multiple core timing modes, supports user-defined timing-related parameters, realizes real-time parsing, storage and retrieval of mode parameters, can complete fast mode switching, adapt to different cleaning scenario requirements, and output parsed mode parameters and switching control signals. The multi-mode timing configuration module includes a core mode preset unit, a parameter parsing and storage unit, and a mode switching unit; The core mode preset unit presets three core timing modes: continuous timing, pulse timing, and stepped timing. The continuous timing mode supports continuous operation from 10s to 180min, with the frequency adaptively adjusted following the resonant frequency and the reference frequency being 28kHz±2%. The pulse timing mode has a minimum pulse period of 100ms. The stepped timing mode supports 3-5 independent parameter configuration steps and outputs preset core timing mode parameters. The parameter parsing and storage unit uses an ARM Cortex-M4 processor to parse and store timing mode parameters in real time. It supports parameter retention even when power is off. It can also remotely modify and batch configure parameters via RS485 bus. Remote communication uses a standard communication protocol and a fixed data frame format, and outputs parsed parameters and storage feedback signals. The mode switching unit adopts a fast response mechanism. The pulse trigger signal is generated by the processor timer module. The step-by-step switching adopts a smooth transition algorithm. It can quickly switch between three core timing modes according to user instructions or cleaning scenario requirements and output mode switching control signals. In this implementation scheme, the three timing modes of the core mode preset unit are designed for different cleaning scenarios. The continuous timing mode is suitable for routine cleaning of small precision parts. The pulse timing mode uses high-frequency pulsed ultrasonic emission and is suitable for cleaning vulnerable and soft parts, avoiding damage to parts caused by prolonged ultrasonic waves. The stepped timing mode allows for the setting of ultrasonic power and frequency parameters in stages according to the cleaning difficulty, and is suitable for gradient cleaning of complex structures and heavily soiled parts. The parameter parsing and storage unit is based on an ARM Cortex-M4 processor to build the hardware control core, and is equipped with 8MB of Flash memory to save parameters after power failure. The RS485 bus communication baud rate is set to 9600bps, adopting the Modbus-RTU standard communication protocol. The data frame format is 1 start bit, 8 data bits, 1 stop bit, and no parity bit. It supports local configuration of a single device and remote batch parameter modification of multiple devices, and can realize a network configuration of up to 32 devices. The processor timer module of the mode switching unit adopts a 16-bit general-purpose timer. The duty cycle of the pulse trigger signal can be customized within the range of 10%-90%. The smooth transition algorithm for step-by-step switching uses linear interpolation to adjust the ultrasonic parameters. The mode switching response time is ≤10ms, ensuring the continuity and stability of the system operation when switching between different cleaning scenarios.

[0022] The multi-source data fusion module is used to collect multi-dimensional operational data during the cleaning process. It performs noise reduction, registration and deviation correction on the collected data to eliminate data noise and deviation interference, integrates it into standardized data, provides reliable data support for subsequent resonance tracking, anomaly detection and closed-loop control, and outputs clean fusion operation data. The multi-source data fusion module includes a multi-dimensional data acquisition unit, a data preprocessing unit, and a data output unit; The multi-dimensional data acquisition unit collects multi-dimensional operational data during the cleaning process, including cleaning fluid temperature, liquid level, transducer impedance, and sound intensity distribution data. The liquid level detection uses a magnetostrictive sensor, and the temperature detection uses a platinum resistance element. The sensors are set with a fixed sampling period and output multi-dimensional raw operational data. The data preprocessing unit uses a 5×5 median filtering algorithm to denoise the raw running data, completes data registration through an iterative nearest point algorithm, and corrects sound field deviations by combining a finite element sound field correction algorithm. The data fusion adopts a reasonable timing and weight allocation strategy to ensure that the parameter transmission delay is ≤50ms. The transmission delay is guaranteed by a caching optimization strategy, and the output is clean fused running data. The data output unit synchronously transmits the cleanroom integration operation data to the resonant frequency adaptive tracking module and the abnormal state detection module, providing data support for subsequent control links and outputting data transmission feedback signals; In this implementation scheme, the magnetostrictive liquid level sensor of the multi-dimensional data acquisition unit has a measurement accuracy of ±0.1mm, and the platinum resistance temperature sensor uses the PT1000 type with a measurement accuracy of ±0.2℃. The unified sampling period of all sensors is set to 100ms. The analog data is converted into 16-bit digital data through the data acquisition card to ensure the accuracy and synchronization of data acquisition. The 5×5 median filtering algorithm of the data preprocessing unit achieves noise reduction of the raw data through a sliding window, effectively filtering out impulse noise and random noise during the sensor acquisition process. The registration iteration error threshold of the iterative nearest point algorithm is set to 0.01 to ensure the spatial dimension matching of the multi-dimensional data. The finite element sound field correction algorithm is based on a preset three-dimensional sound field model to correct the deviation of the actual acquired sound intensity distribution data. The timing strategy of data fusion adopts alignment according to the acquisition timestamp, and the weight allocation is set according to the sensor measurement accuracy. The higher the accuracy, the larger the weight ratio. After processing, the parameter transmission delay is stabilized at 30-50ms, providing highly reliable standardized data for subsequent modules. The data output unit adopts a dual data transmission channel to synchronously transmit cleanroom fusion operation data to the resonant frequency adaptive tracking module and the abnormal state detection module, respectively. Each frame of data is accompanied by a check code to ensure that there is no loss or error in the data transmission process. At the same time, the data transmission status is fed back in real time, and a retransmission mechanism is triggered in time when there is an abnormality.

[0023] The resonant frequency adaptive tracking module is used to monitor the resonant state of the transducer in real time, dynamically adjust the driving frequency to compensate for load changes during the cleaning process, maintain the resonant stability of the transducer, ensure the acoustic energy transmission efficiency, and output the adjusted driving frequency parameters and resonant state monitoring data. The resonant frequency adaptive tracking module includes a resonant state monitoring unit, a drive frequency adjustment unit, and a temperature compensation unit. The resonant state monitoring unit is based on a DSP processor and monitors the resonant state of the transducer in real time. It uses the impedance phase angle as the core monitoring index and outputs the transducer resonant state data. The drive frequency adjustment unit adopts the gradient approximation method, sets reasonable step size and convergence conditions, and dynamically adjusts the drive frequency to compensate for load changes based on the transducer resonance state data. Load compensation is completed within 0.3s, ensuring that the resonance point deviation is ≤±100Hz. The drive circuit adopts a full-bridge topology and soft-switching technology. The power MOSFET uses third-generation semiconductor materials, and the overall board efficiency is ≥92%. The soft-switching technology sets the corresponding dead time and drive strategy, and outputs the adjusted drive frequency parameters. The temperature compensation unit performs temperature coefficient compensation every 15 minutes. It uses a calibrated mathematical model to correct the influence of temperature changes on the resonance state, ensuring that the sound energy transmission efficiency is ≥90% and outputting a temperature compensation signal. In this implementation scheme, the DSP processor of the resonant state monitoring unit is a TMS320F28335, with a sampling frequency set to 1MHz. It collects the impedance phase angle data of the transducer in real time, with a phase angle detection accuracy of ±0.1°. When the phase angle deviates from the resonant point phase angle by ±5°, a drive frequency adjustment command is triggered. The gradient approximation method of the drive frequency adjustment unit is set with an initial step size of 50Hz. When the resonant point deviation decreases to ±200Hz, the step size is automatically adjusted to 10Hz. The convergence condition is that the resonant point deviation is ≤±100Hz, ensuring that load compensation is completed within 0.3s. The full-bridge topology of the drive circuit uses IGBT power transistors, and the third-generation semiconductor material is silicon carbide. The dead time is set to 2μs, effectively avoiding bridge arm shoot-through. The actual operating efficiency of the entire board can reach 92%-95%, significantly improving the energy efficiency ratio of sound energy transmission. The calibration mathematical model of the temperature compensation unit is a linear fitting model, which is constructed based on the measured data of the resonant frequency at different temperatures. The cleaning fluid temperature data is automatically collected every 15 minutes, and the compensation value is calculated and the driving frequency is adjusted by substituting it into the model. Even under the condition of large changes in the cleaning fluid temperature, the sound energy transmission efficiency can still be stably maintained at 90%-95%.

[0024] The abnormal state detection module is used to construct a baseline feature space for cleaning parameters. Combined with a preset abnormal judgment threshold, it identifies abnormal situations during the operation of the cleaning equipment, promptly captures abnormal fluctuations in temperature, liquid level, and impedance parameters, and outputs abnormal detection results and alarm trigger signals. The abnormal state detection module includes a baseline feature space construction unit, an abnormal identification unit, and an abnormal reporting unit; The benchmark feature space construction unit is based on the K-means clustering algorithm. It uses sufficient normal working condition data as the training sample set. The number of clusters is determined by the elbow rule. The clean parameter benchmark feature space is constructed. The dimensionality is reduced by principal component analysis to retain ≥95% information entropy. The principal component selection is based on the eigenvalue setting. The benchmark feature space parameters are output. The anomaly identification unit has preset anomaly judgment thresholds, including temperature gradient > 3℃ / min, liquid level fluctuation > 5mm, and impedance change > 15%. These anomaly thresholds are calibrated through multiple experiments and combined with the benchmark feature space parameters to identify equipment operation anomalies and output anomaly detection results. After identifying an anomaly, the anomaly reporting unit reports the anomaly event simultaneously through local audible and visual alarms and remote communication. The alarm and reporting settings include fixed response delays and fault code rules, and the unit outputs an anomaly reporting signal. In this implementation scheme, the K-means clustering algorithm of the baseline feature space construction unit selects 10,000 sets of parameter data such as temperature, liquid level, impedance, and sound intensity under normal operating conditions as the training sample set. The elbow rule is used to determine the number of clusters to be 8. Principal component analysis is used to select the first 3 principal components, and the information entropy retention rate can reach 95%-97%, effectively reducing the data dimensionality and improving the efficiency of subsequent anomaly identification. The anomaly judgment threshold of the anomaly identification unit has been calibrated through more than 500 experiments under different operating conditions. In addition to the core thresholds of temperature gradient > 3℃ / min, liquid level fluctuation > 5mm, and impedance change > 15%, an auxiliary judgment threshold of sound intensity distribution deviation > 10% is also set. A multi-parameter fusion judgment method is adopted to avoid misjudgment caused by fluctuation of a single parameter. The accuracy of anomaly identification can reach more than 98%. The local audible and visual alarm of the anomaly reporting unit uses a red LED warning light and a buzzer. The warning light flashes at a frequency of 2Hz, and the buzzer sounds at a frequency of 1kHz. The remote communication adopts a 4G / Ethernet dual-mode communication method. The fault code is compiled according to the rule of "fault type + fault location + fault time". The response delay is ≤1s, ensuring that maintenance personnel can obtain equipment anomaly information in a timely manner and quickly carry out fault handling.

[0025] The intelligent closed-loop control module is used to link the start-stop control module when an abnormal detection signal is received, dynamically adjust the system operating parameters, balance the cleaning effect and energy consumption, ensure that the equipment can quickly resume stable operation after the abnormality is handled, and output closed-loop control commands and optimized operating parameters. The intelligent closed-loop control module includes an optimization target setting unit, an operating parameter adjustment unit, and a stability recovery unit; The optimization target setting unit adopts the model predictive control algorithm, with the goal of "maximum cleaning efficiency and minimum energy consumption", establishes a target function with unified dimensions, sets reasonable weight coefficients for the target function, and outputs the optimized target parameters; The operating parameter adjustment unit adopts a rolling time domain optimization strategy, sets the corresponding prediction step size and control cycle, and dynamically adjusts the system amplitude, frequency and running time. With the help of an electric regulating valve and a water pump with a response time of <50ms, the liquid level adjustment is completed within 200ms. The liquid level adjustment adopts a PID algorithm and outputs the adjusted operating parameters and control commands. After the anomaly handling is completed, the stability recovery unit ensures that the system returns to stable operation within 3 seconds. The stable operation is defined by clear judgment criteria and outputs a system stable operation feedback signal. In this implementation scheme, the model predictive control algorithm of the target setting unit is built based on MATLAB's MPC toolbox. The weighting coefficient of cleaning efficiency in the objective function is set to 0.6, and the weighting coefficient of energy consumption is set to 0.4, achieving an optimal balance between cleaning effect and energy consumption. Simultaneously, a lower limit for cleaning efficiency and an upper limit for energy consumption are set to avoid over-optimization of a single objective. The rolling time-domain optimization strategy of the operating parameter adjustment unit is set with a prediction step size of 10 steps and a control cycle of 200ms. Both the electric regulating valve and the water pump adopt servo control, with a response time controlled within 30-50ms. The PID algorithm for liquid level adjustment adopts incremental PID, and the proportional coefficient, integral coefficient, and derivative coefficient are determined through on-site debugging and optimization. The liquid level adjustment accuracy can reach ±0.5mm within 200ms, enabling rapid compensation for liquid level fluctuations. The system stability judgment criteria for the stable recovery unit are: temperature fluctuation ≤ ±0.5℃, liquid level fluctuation ≤ ±1mm, impedance fluctuation ≤ ±2%, and resonant point deviation ≤ ±100Hz. The above parameters are maintained stably for more than 500ms. After the abnormality is handled, the system gradually restores the operating parameters to ensure that it reaches a stable operating state within 3s, avoiding equipment shock caused by sudden parameter changes.

[0026] It also includes a cleaning tank and a transducer structure module, which includes a cleaning tank structure unit and a transducer unit. The cleaning tank structural unit is made of ZL114A aluminum alloy in one piece and is manufactured by die casting. Microchannels are embedded in the inner wall to enhance the heat exchange effect. The microchannels are set with fixed cross-sectional dimensions and spacing. The connection is a sealed joint with IP68 protection level, which is suitable for wide temperature conditions from -40℃ to 125℃. Output cleaning tank structural parameters and protection performance parameters. The transducer unit uses lead scandate-lead titanate piezoelectric material. The composition ratio, preparation process and polarization conditions of the piezoelectric material are clearly defined. The transducer has an electroacoustic conversion efficiency of ≥95% after continuous operation at 80℃ for 2000 hours. The electroacoustic conversion efficiency and protection performance are verified by corresponding standard tests, and the transducer performance parameters are output. In this implementation scheme, the ZL114A aluminum alloy die-casting process of the cleaning tank structural unit adopts high-pressure die casting with a die-casting pressure of 120MPa and a mold temperature controlled at 200-220℃. The microchannel cross-section of the inner wall of the cleaning tank is rectangular with dimensions of 2mm×3mm and a channel spacing of 10mm. Circulating cooling water is introduced into the microchannels for heat exchange. The IP68 protection level sealing joints at the connection points are made of fluororubber. After high and low temperature cycle testing, there is no leakage or deformation under a wide temperature range of -40℃ to 125℃. The lead scandate-lead titanate piezoelectric material of the transducer unit has a Sc doping ratio of 6%. The ceramic powder is prepared by the sol-gel method and sintered and polarized at 1200℃. The piezoelectric constant d33 ≥ 500pC / N. The transducer is fixed to the bottom of the cleaning tank by adhesive bonding. After a constant temperature aging test at 80℃, the electroacoustic conversion efficiency remains at 95%-96% after 2000 hours of continuous operation, with no performance degradation.

[0027] It also includes a cleaning fluid management module, which includes a conductivity monitoring unit, a fluid replenishment and compensation unit, and a dead-angle cleaning unit. The conductivity monitoring unit uses a four-electrode conductivity sensor with a fixed measurement range and accuracy to monitor the conductivity of the cleaning fluid in real time. When the conductivity is >50μS / cm, the filtration device is automatically started. The filtration device is set with a fixed filtration cycle and a stop threshold, and outputs conductivity monitoring data and filtration control signals. The liquid replenishment unit has a built-in liquid replenishment device that is linked to the liquid level sensor to realize automatic replenishment of cleaning fluid. It can set the corresponding liquid replenishment threshold and speed, and output liquid replenishment control signal and liquid level monitoring data. The dead-angle-free cleaning unit uses a low-viscosity synthetic heat-conducting fluid. The viscosity, thermal conductivity and applicable temperature range of the heat-conducting fluid are clearly defined. It works with a circumferential emission transducer to achieve 360° dead-angle-free cleaning. The circumferential emission transducers are arranged in a fixed number and at a fixed angle to output cleaning control parameters. In this implementation scheme, the four-electrode conductivity sensor of the conductivity monitoring unit has a measurement range of 0-200 μS / cm and a measurement accuracy of ±1%FS. The filtration device uses a precision filter element with a filtration accuracy of 5 μm and a filtration cycle of 30 min. Filtration automatically stops when the conductivity of the cleaning fluid drops to 30 μS / cm, achieving circulation and purification of the cleaning fluid. The replenishment device of the replenishment unit uses a peristaltic pump, and the replenishment rate can be adjusted within the range of 5-50 mL / s. The replenishment threshold is set to start replenishment when the liquid level is 2 mm below the set value and stop replenishing when the set value is reached. The liquid level monitoring data interacts with the multi-source data fusion module in real time to ensure the accuracy of replenishment. The low-viscosity synthetic thermally conductive fluid of the no-dead-angle cleaning unit has a viscosity of 5 mPa·s (25℃) and a thermal conductivity of 0.5 W / (m·K). The applicable temperature range is -20℃ to 100℃. The circumferential transmitting transducers are arranged in a circular array, with 8 transducers in each circle and an angle of 45° between adjacent transducers. A total of 2-3 circles are set up to achieve 360° ultrasonic coverage without dead angles in the cleaning tank, thoroughly removing blind holes and crevices of complex structural parts.

[0028] It also includes a protection and energy-saving operation module, which includes an EMC protection unit, a data storage unit, and an energy-saving control unit; The EMC protection unit adopts an industrial-grade EMC protection design, including power filtering, signal shielding and grounding, to ensure the system's anti-interference capability and output EMC protection parameters. The data storage unit is equipped with an edge computing module with a processing latency of ≤100ms. It supports local storage of 1,000 sets of cleaning logs and parameters. The edge computing module uses an appropriate processor and storage medium to output data storage feedback signals. The energy-saving control unit sets clear energy-saving mode switching conditions and power reduction ratios, automatically switches to energy-saving mode during low system load periods, making the system suitable for cleaning in multiple scenarios, with a cleaning qualification rate of ≥97%, which has been verified by experiments, and outputs energy-saving control commands and operation feedback signals. In this implementation scheme, the power supply filtering of the EMC protection unit adopts a three-stage EMI filter to effectively suppress conducted interference in the power grid. The signal shielding uses a metal shield and grounding, with a grounding resistance ≤4Ω. EMC testing shows that the system's radiated immunity meets the GB / T17626.3-2016 level 3 standard, and its conducted immunity meets the GB / T17626.4-2018 level 4 standard, making it adaptable to the complex electromagnetic environment of industrial sites. The edge computing module of the data storage unit uses an STM32H743 processor and is equipped with a 128MB SD card for local storage. The processing latency is controlled within 80-100ms. It supports local querying, USB export, and remote uploading of cleaning logs and parameters. The 1000-set storage capacity can meet the needs of more than 3 months of regular cleaning data recording. The energy-saving control unit's energy-saving mode switching conditions are set as follows: system load rate <30% for 5 minutes. In energy-saving mode, ultrasonic power is reduced by 50%, and the drive frequency remains at the resonant frequency. According to actual tests, the system energy consumption is reduced by 45%-50% in energy-saving mode, while the cleaning qualification rate is still maintained at 97%-99%. It significantly reduces energy consumption in low-load cleaning scenarios while meeting cleaning quality requirements.

[0029] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium is implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0030] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention should all be covered within the scope of the claims of the present invention.

Claims

1. A multi-timing mode ultrasonic cleaning intelligent control system, characterized in that, It includes a cleaning condition modeling module, a multi-mode timing configuration module, a multi-source data fusion module, a resonant frequency adaptive tracking module, an abnormal state detection module, and an intelligent closed-loop control module. The cleaning condition modeling module is used to acquire transducer vibration field data, construct a three-dimensional sound field model in combination with relevant algorithms, optimize the transducer layout, ensure the uniformity of the sound field in the cleaning tank, provide a basic sound field model support for subsequent cleaning control, and output the optimized three-dimensional sound field model and transducer layout parameters. The multi-mode timing configuration module is used to preset multiple core timing modes, supports user-defined timing-related parameters, realizes real-time parsing, storage and retrieval of mode parameters, can complete rapid mode switching, adapt to different cleaning scenario requirements, and output parsed mode parameters and switching control signals. The multi-source data fusion module is used to collect multi-dimensional operational data during the cleaning process, perform noise reduction, registration and deviation correction on the collected data, eliminate data noise and deviation interference, integrate into standardized data, provide reliable data support for subsequent resonance tracking, anomaly detection and closed-loop control, and output clean fusion operation data. The resonant frequency adaptive tracking module is used to monitor the resonant state of the transducer in real time, dynamically adjust the driving frequency to compensate for load changes during the cleaning process, maintain the resonant stability of the transducer, ensure the acoustic energy transmission efficiency, and output the adjusted driving frequency parameters and resonant state monitoring data. The abnormal state detection module is used to construct a baseline feature space for cleaning parameters, and combined with a preset abnormal judgment threshold, identify abnormal situations during the operation of the cleaning equipment, promptly capture abnormal fluctuations in temperature, liquid level, and impedance parameters, and output abnormal detection results and alarm trigger signals. The intelligent closed-loop control module is used to link the start-stop control module when receiving an anomaly detection signal, dynamically adjust the system operating parameters, balance the cleaning effect and energy consumption, ensure that the equipment can quickly resume stable operation after the anomaly is handled, and output closed-loop control commands and optimized operating parameters.

2. The intelligent control system for ultrasonic cleaning with multiple timing modes according to claim 1, characterized in that: The cleaning condition modeling module includes a vibration field data acquisition unit, a three-dimensional sound field modeling unit, and a transducer layout optimization unit. The vibration field data acquisition unit acquires transducer vibration field data through a laser Doppler vibration meter and an impedance sensor array. Simultaneously, a PVDF sound intensity sensor is used to set sound intensity monitoring points throughout the cleaning tank. The sensor has a built-in temperature compensation circuit, which can automatically correct the influence of ambient temperature (0-80℃) on the measurement results, ensuring measurement accuracy ≤±2%. The temperature compensation circuit is calibrated experimentally to determine the compensation coefficient and outputs the original vibration field data and sound intensity monitoring data. The three-dimensional sound field modeling unit combines the finite element analysis method to analyze and process the collected vibration field data and sound intensity monitoring data, construct a three-dimensional sound field model, and use mesh adaptive technology to densify the edge mesh of the cleaning tank. The mesh densification criterion is to start when the edge sound field gradient exceeds the preset range, and output the three-dimensional sound field model. The transducer layout optimization unit introduces a genetic algorithm, sets reasonable fitness functions, constraints and iteration parameters, and iteratively optimizes the installation angle and spacing of the transducers to make the sound field uniformity error ≤8% and the cavitation effect coverage ≥95%. The sound field uniformity error and cavitation effect coverage are verified by corresponding test methods, and the optimized transducer layout parameters are output.

3. The intelligent control system for ultrasonic cleaning with multiple timing modes according to claim 1, characterized in that: The multi-mode timing configuration module includes a core mode preset unit, a parameter parsing and storage unit, and a mode switching unit; The core mode preset unit presets three core timing modes: continuous timing, pulse timing, and stepped timing. The continuous timing mode supports continuous operation from 10s to 180min, with the frequency adaptively adjusted following the resonant frequency and the reference frequency being 28kHz±2%. The pulse timing mode has a minimum pulse period of 100ms. The stepped timing mode supports 3-5 independent parameter configuration steps and outputs preset core timing mode parameters. The parameter parsing and storage unit uses an ARM Cortex-M4 processor to parse and store the timing mode parameters in real time. It supports parameter retention even when power is off. At the same time, it can remotely modify and batch configure parameters via RS485 bus. Remote communication adopts a standard communication protocol and a fixed data frame format, and outputs the parsed parameters and storage feedback signals. The mode switching unit adopts a fast response mechanism. The pulse trigger signal is generated by the processor timer module, and the step-by-step switching adopts a smooth transition algorithm. It can quickly switch between three core timing modes according to user instructions or cleaning scenario requirements, and output mode switching control signals.

4. The intelligent control system for ultrasonic cleaning with multiple timing modes according to claim 1, characterized in that: The multi-source data fusion module includes a multi-dimensional data acquisition unit, a data preprocessing unit, and a data output unit; The multi-dimensional data acquisition unit collects multi-dimensional operational data during the cleaning process, including cleaning fluid temperature, liquid level, transducer impedance, and sound intensity distribution data. The liquid level detection uses a magnetostrictive sensor, and the temperature detection uses a platinum resistance element. The sensors are set with a fixed sampling period and output multi-dimensional raw operational data. The data preprocessing unit uses a 5×5 median filtering algorithm to denoise the raw running data, completes data registration through an iterative nearest point algorithm, and corrects sound field deviations by combining a finite element sound field correction algorithm. The data fusion adopts a reasonable timing and weight allocation strategy to ensure that the parameter transmission delay is ≤50ms. The transmission delay is guaranteed by a caching optimization strategy, and the output is clean fused running data. The data output unit synchronously transmits the cleanroom integration operation data to the resonant frequency adaptive tracking module and the abnormal state detection module, providing data support for subsequent control links and outputting data transmission feedback signals.

5. The intelligent control system for ultrasonic cleaning with multiple timing modes according to claim 1, characterized in that: The resonant frequency adaptive tracking module includes a resonant state monitoring unit, a drive frequency adjustment unit, and a temperature compensation unit. The resonant state monitoring unit is based on a DSP processor and monitors the resonant state of the transducer in real time. It uses the impedance phase angle as the core monitoring index and outputs the transducer resonant state data. The drive frequency adjustment unit adopts the gradient approximation method, sets reasonable step size and convergence conditions, and dynamically adjusts the drive frequency to compensate for load changes based on the transducer resonance state data. Load compensation is completed within 0.3s, ensuring that the resonance point deviation is ≤±100Hz. The drive circuit adopts a full-bridge topology and soft-switching technology. The power MOSFET uses third-generation semiconductor materials, and the overall board efficiency is ≥92%. The soft-switching technology sets the corresponding dead time and drive strategy, and outputs the adjusted drive frequency parameters. The temperature compensation unit performs temperature coefficient compensation every 15 minutes, using a calibrated mathematical model to correct the influence of temperature changes on the resonance state, ensuring that the sound energy transmission efficiency is ≥90%, and outputs a temperature compensation signal.

6. The intelligent control system for ultrasonic cleaning with multiple timing modes according to claim 1, characterized in that: The abnormal state detection module includes a baseline feature space construction unit, an abnormal identification unit, and an abnormal reporting unit. The benchmark feature space construction unit is based on the K-means clustering algorithm, using sufficient normal working condition data as the training sample set. The number of clusters is determined by the elbow rule. A clean parameter benchmark feature space is constructed, and dimensionality reduction is performed by principal component analysis to retain ≥95% information entropy. The principal component selection is based on the eigenvalue setting, and the benchmark feature space parameters are output. The anomaly identification unit presets anomaly judgment thresholds, including temperature gradient > 3℃ / min, liquid level fluctuation > 5mm, and impedance change > 15%. These anomaly thresholds are calibrated through multiple experiments and combined with the benchmark feature space parameters to identify equipment malfunctions and output anomaly detection results. After identifying an anomaly, the anomaly reporting unit reports the anomaly event synchronously through local audible and visual alarms and remote communication. The alarm and reporting settings include fixed response delays and fault code rules, and the unit outputs an anomaly reporting signal.

7. The intelligent control system for ultrasonic cleaning with multiple timing modes according to claim 1, characterized in that: The intelligent closed-loop control module includes an optimization target setting unit, an operating parameter adjustment unit, and a stability recovery unit; The optimization target setting unit adopts a model predictive control algorithm, with the goal of "maximum cleaning efficiency and minimum energy consumption", establishes a target function with unified dimensions, sets reasonable weight coefficients for the target function, and outputs the optimization target parameters; The operating parameter adjustment unit adopts a rolling time-domain optimization strategy, sets the corresponding prediction step size and control cycle, and dynamically adjusts the system amplitude, frequency and running time. With the help of an electric regulating valve and a water pump with a response time of <50ms, the liquid level adjustment is completed within 200ms. The liquid level adjustment adopts a PID algorithm and outputs the adjusted operating parameters and control commands. After the anomaly is handled, the stability recovery unit ensures that the system returns to stable operation within 3 seconds. The stable operation is defined by clear judgment criteria and outputs a system stable operation feedback signal.

8. The intelligent control system for ultrasonic cleaning with multiple timing modes according to claim 1, characterized in that: It also includes a cleaning tank and a transducer structure module, which includes a cleaning tank structure unit and a transducer unit. The cleaning tank structural unit is made of ZL114A aluminum alloy in one piece and is prepared by die casting process. Microchannels are embedded in the inner wall to enhance the heat exchange effect. The microchannels are set with fixed cross-sectional size and spacing. The connection is a sealed joint with IP68 protection level, which is suitable for wide temperature conditions of -40℃ to 125℃. The cleaning tank structural parameters and protection performance parameters are output. The transducer unit uses lead scandate-lead titanate piezoelectric material. The composition ratio, preparation process and polarization conditions of the piezoelectric material are specified. The transducer has an electroacoustic conversion efficiency of ≥95% after continuous operation at 80℃ for 2000 hours. The electroacoustic conversion efficiency and protection performance are verified by corresponding standard tests, and the transducer performance parameters are output.

9. The intelligent control system for ultrasonic cleaning with multiple timing modes according to claim 1, characterized in that: It also includes a cleaning fluid management module, which includes a conductivity monitoring unit, a fluid replenishment and compensation unit, and a dead-angle-free cleaning unit. The conductivity monitoring unit uses a four-electrode conductivity sensor with a fixed measurement range and accuracy to monitor the conductivity of the cleaning fluid in real time. When the conductivity is >50μS / cm, the filtration device is automatically started. The filtration device is set with a fixed filtration cycle and a stop threshold, and outputs conductivity monitoring data and filtration control signals. The liquid replenishment unit has a built-in liquid replenishment device, which is linked to the liquid level sensor to realize automatic compensation of cleaning fluid. It can set the corresponding liquid replenishment threshold and speed, and output liquid replenishment control signal and liquid level monitoring data. The no-dead-angle cleaning unit uses a low-viscosity synthetic heat-conducting fluid, with clearly defined viscosity, thermal conductivity and applicable temperature range. It works in conjunction with circumferential emission transducers to achieve 360° no-dead-angle cleaning. The circumferential emission transducers are arranged in a fixed number and at a fixed angle to output cleaning control parameters.

10. The intelligent control system for ultrasonic cleaning with multiple timing modes according to claim 1, characterized in that: It also includes a protection and energy-saving operation module, which includes an EMC protection unit, a data storage unit, and an energy-saving control unit; The EMC protection unit adopts an industrial-grade EMC protection design, including power filtering, signal shielding and grounding treatment, to ensure the system's anti-interference capability and output EMC protection parameters. The data storage unit is equipped with an edge computing module with a processing latency of ≤100ms. It supports local storage of 1,000 sets of cleaning logs and parameters. The edge computing module uses an appropriate processor and storage medium to output data storage feedback signals. The energy-saving control unit sets clear energy-saving mode switching conditions and power reduction ratios, automatically switching to energy-saving mode during low-load periods, making the system suitable for cleaning in multiple scenarios, with a cleaning qualification rate of ≥97%, which is verified by experiments, and outputs energy-saving control commands and operation feedback signals.

Images