Safe operation protection method for ozone power supply system

Abstract

The invention relates to a safe operation protection method for an ozone power supply system. The method comprises the following steps: constructing an insulation safety evaluation model, generating a multi-stage safety threshold system, executing a graded protection action, and generating an insulation fault positioning report; setting a reflux judgment condition, starting a bidirectional energy feedback and absorption mechanism, continuously monitoring a reflux state, controlling the output power of a power supply, cutting off a main circuit switch, prohibiting the restart of the power supply, and recording reflux related operation data; constructing a thermal runaway pre-judgment model, presetting a thermal risk judgment criterion, triggering a thermal runaway early warning mechanism, and starting a hierarchical thermal protection strategy; setting an electromagnetic interference safety threshold, triggering an interference suppression response, adopting a filtering adjustment mechanism, adjusting the working frequency of a power supply, generating an electromagnetic interference standard exceeding fault report, and recording related data. Multi-dimensional intelligent protection is achieved, the safety and reliability of equipment are improved, the energy utilization rate is increased, operation and maintenance convenience is enhanced, and the service life of the equipment is prolonged.

Description

Technical Field

[0001] This invention belongs to the field of ozone power supply safety control technology, specifically relating to a method for protecting the safe operation of an ozone power supply system. Background Technology

[0002] Ozone generators are core equipment in environmental protection, water treatment, and air purification, and their operational stability and safety are directly determined by the supporting ozone power supply system. This power supply needs to operate under complex conditions of high voltage, high frequency, and high power for extended periods. The load side exhibits nonlinear characteristics such as dielectric barrier discharge and corona discharge, and is prone to multiple risks during operation, including insulation degradation, energy backflow, heat accumulation, and electromagnetic interference. This places stringent requirements on the reliability and adaptability of safety protection technologies.

[0003] With the upgrading of the environmental protection industry and the development of intelligent equipment, high-power ozone generators with long continuous operating times are being used more and more widely. The market's requirements for the active protection capabilities, dynamic adaptability, and ease of operation and maintenance of power supply systems are constantly increasing. Although existing safety protection schemes can achieve basic risk prevention and control, they have obvious limitations: they are difficult to adapt to the dynamic operating condition changes throughout the entire life cycle of the power supply, and their ability to predict risks such as insulation degradation and thermal runaway is insufficient; energy backflow handling is mostly based on passive absorption, without taking into account energy recovery and grid friendliness; electromagnetic interference suppression mostly adopts fixed parameter schemes, and the dynamic adjustment effect is limited; and there is a lack of a sound fault data tracing mechanism, which is not conducive to operation and maintenance optimization.

[0004] Meanwhile, the increasing demands for industrial automation and energy conservation further require ozone power supply protection technology to possess characteristics such as graded response, intelligent early warning, and high energy efficiency. Against this backdrop, there is a need to develop a safe operation protection method that covers multiple core risks, adapts to complex operating conditions, and possesses active protection and intelligent traceability capabilities. This is to meet the development needs of ozone power supply systems for high reliability, long lifespan, and intelligence, and to provide comprehensive technical support for the stable and efficient operation of ozone generators. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, this invention provides a method for protecting the safe operation of an ozone power supply system.

[0006] The objective of this invention can be achieved through the following technical solution: a method for protecting the safe operation of an ozone power supply system, comprising: S1: Real-time acquisition of insulation status parameters of the main circuit, construction of insulation safety assessment model, generation of multi-level safety threshold system, execution of graded protection actions according to the threshold range of the insulation status parameters, and generation of insulation fault location report; S2: Real-time acquisition of energy backflow signal from the ozone power supply load side, setting backflow judgment conditions, starting bidirectional energy feedback and absorption mechanism, continuously monitoring backflow status, controlling power output power, cutting off main circuit switch, prohibiting power restart and recording backflow-related operating data; S3: Obtain the temperature parameters of the ozone power module, build a thermal runaway prediction model based on historical data, preset thermal risk judgment criteria, trigger the thermal runaway early warning mechanism, and start a graded thermal protection strategy. S4: By acquiring the electromagnetic interference signal from the ozone power supply output side in real time, setting the electromagnetic interference safety threshold, and triggering the interference suppression response based on the electromagnetic interference signal, the power supply operating frequency is first adjusted using a filtering adjustment mechanism, generating an electromagnetic interference exceeding the standard fault report and simultaneously recording the relevant exceeding data.

[0007] Specifically, the construction process of the insulation safety assessment model is as follows: First, obtain the insulation operation status records and fault cases throughout the entire life cycle of the power supply. Combine the insulation failure mechanism analysis to screen the characteristic indicators that affect safety. Calculate the correlation weight of the characteristic indicators with insulation safety through statistical modeling. Then, build the core operation framework of the model using a weighted comprehensive evaluation mechanism and simultaneously integrate the insulation characteristic difference data under different operating conditions for dynamic calibration.

[0008] Specifically, the multi-level safety threshold system is generated as follows: statistically clustering historical insulation fault data, screening influencing factors and calculating weights, dynamically correcting the threshold boundary by simulating the insulation failure evolution process, and continuously obtaining fault data to iteratively optimize the threshold range.

[0009] Specifically, the specific generation process of the insulation fault location report is as follows: real-time acquisition of insulation anomaly parameters, feature extraction, construction of a feature vector set, similarity calculation with the multi-dimensional fault feature matrix in the preset fault template library, simultaneous application of time series analysis mechanism to dynamically model the fault evolution process, and based on the coupling analysis of similarity calculation results and trend prediction model, the fault location is locked, and finally, an insulation fault location research report is generated.

[0010] Specifically, the backflow determination conditions include: based on the energy backflow signal, firstly, synchronously acquire the instantaneous values ​​of load-side voltage and current, and through a data fusion mechanism, calculate the instantaneous power direction and amplitude change. During the backflow determination stage, calculate the deviation rate between the real-time calculation results and the preset benchmark value, and simultaneously monitor the duration of the deviation.

[0011] Specifically, the bidirectional energy feedback and absorption mechanism includes: firstly, using a real-time grid parameter monitoring algorithm to quantitatively assess the grid's dynamic acceptance capacity; based on the assessment results, triggering the bidirectional energy feedback and absorption mechanism to safely inject system backflow energy into the grid; and activating an energy caching strategy to temporarily store excess backflow energy in an energy storage unit while continuously monitoring grid parameters.

[0012] Specifically, the thermal runaway prediction model is constructed as follows: historical temperature time series data and thermal failure cases of the ozone power module are obtained, key temperature characteristic parameters of thermal runaway are screened by combining heat conduction mechanism analysis, the correlation strength between key temperature characteristic parameters of thermal runaway and thermal runaway risk is calculated by statistical modeling, and then the core computing framework is built by using a trend extrapolation mechanism, and thermal characteristic difference data under different operating conditions are incorporated for calibration.

[0013] Specifically, the thermal risk determination criteria include: combining the thermal conduction characteristics of the power module with the thermal runaway evolution law, integrating real-time temperature values, temperature rise rate, and temperature fluctuation amplitude as determination indicators, assigning judgment weights to the indicators according to the degree of thermal failure impact, performing thermal risk status determination through a multi-indicator linkage verification mechanism, and continuously iterating and correcting indicator weights and judgment thresholds based on newly acquired thermal failure cases and temperature monitoring data.

[0014] Specifically, the thermal runaway early warning mechanism includes: first, structurally encapsulating the early warning information, simultaneously triggering local warning logic and remote data transmission process, and establishing an early warning information traceability index to associate the temperature parameters, the thermal risk level assessment results, and historical cases.

[0015] Specifically, the graded thermal protection strategy includes: simultaneously performing heat dissipation regulation and gradient power reduction operations, acquiring temperature data in real time to check the cooling effect, executing a power supply cut-off process based on the cooling effect, and gradually dissipating residual energy by utilizing energy release before the cut-off.

[0016] Specifically, the filtering adjustment mechanism includes: first, adjusting the suppression parameters based on the electromagnetic interference exceeding the standard frequency band; monitoring the changes in the interference signal after filtering in real time; dynamically optimizing the power supply operating frequency according to the suppression effect; and maintaining the stable state of output voltage and power through collaborative control logic.

[0017] Specifically, the electromagnetic interference exceeding the standard fault report generation process is as follows: based on the actual suppression effect after filter adjustment and power supply operating frequency adjustment, integrate electromagnetic interference characteristic parameters and corresponding filter parameters and frequency adjustment trajectory, synchronously associate with the current operating condition information of the power supply, classify the recorded exceeding data according to the operating condition type to establish a retrieval index, and immediately synchronize it to the system monitoring terminal after generation.

[0018] This invention provides comprehensive protection against the core operational risks of ozone power systems through a multi-dimensional, intelligent, and hierarchical safety protection design, and has the following significant beneficial effects: Enhancing safety and reliability: Through an insulation safety assessment model and a multi-level threshold system, dynamic monitoring and precise graded protection of insulation status are achieved, and the fault location mechanism quickly pinpoints the problem point; a thermal runaway prediction model and multi-dimensional judgment criteria provide early warning of thermal risks, and a graded thermal protection strategy avoids secondary faults; a multi-condition judgment and bidirectional feedback mechanism for energy backflow prevents circuit damage caused by backflow; dynamic filtering and frequency adjustment of electromagnetic interference effectively suppress the impact of interference, comprehensively covering the four core risks and significantly reducing the incidence of equipment failures and safety accidents.

[0019] Achieving efficient energy utilization and energy saving: The bidirectional energy feedback and absorption mechanism can not only safely handle the backflow of energy on the load side, but also feed excess backflow energy back to the grid or temporarily store it in the energy storage unit, avoiding the energy waste caused by traditional passive absorption and improving energy utilization. At the same time, the coordinated control logic in the hierarchical thermal protection and electromagnetic interference suppression ensures stable power output while reducing ineffective energy consumption, which meets the industrial demand for energy conservation and emission reduction.

[0020] Enhance the intelligence and convenience of operation and maintenance: Structured documents such as insulation fault location reports and electromagnetic interference exceeding the standard fault reports, combined with complete operation data records and retrieval indexes, provide clear evidence for fault tracing; local warnings and remote transmission of early warning information, as well as the function of associating historical cases, help operation and maintenance personnel quickly troubleshoot problems; dynamic evaluation models supported by full life cycle data facilitate subsequent optimization of protection strategies and significantly reduce maintenance costs and downtime.

[0021] Extending equipment lifespan and operational stability: Proactive prediction and graded protection against issues such as insulation degradation and heat accumulation reduce wear and tear on core components; gradual residual energy release and grid-friendly design prevent the impact of sudden operating conditions on the power system; precise suppression of electromagnetic interference ensures stable coordination between the control circuit and peripheral equipment, comprehensively improving the long-term continuous operation capability of the ozone power system and extending the overall lifespan of the equipment.

[0022] Adaptable to complex operating conditions and scenario expansion needs: The model construction process incorporates characteristic difference data of different operating conditions, multi-level protection strategies and dynamic adjustment mechanisms (such as filter parameters and operating frequency optimization) to enable the system to adapt to complex operating conditions such as high voltage, high frequency, and high power, as well as state changes throughout the entire life cycle; the intelligent and modular protection design can be adapted to ozone generators of different power specifications to meet the application expansion needs of multiple fields such as environmental protection and water treatment. Attached Figure Description

[0023] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0024] Figure 1 This is a flowchart of a method for protecting the safe operation of an ozone power supply system according to the present invention; Figure 2 This is a schematic diagram of a method for protecting the safe operation of an ozone power supply system according to the present invention. Detailed Implementation

[0025] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0026] Please see Figures 1-2 A method for protecting the safe operation of an ozone power supply system, comprising: S1: Real-time acquisition of insulation status parameters of the main circuit, construction of insulation safety assessment model, generation of multi-level safety threshold system, execution of graded protection actions according to the threshold range of the insulation status parameters, and generation of insulation fault location report; S2: Real-time acquisition of energy backflow signal from the ozone power supply load side, setting backflow judgment conditions, starting bidirectional energy feedback and absorption mechanism, continuously monitoring backflow status, controlling power output power, cutting off main circuit switch, prohibiting power restart and recording backflow-related operating data; S3: Obtain the temperature parameters of the ozone power module, build a thermal runaway prediction model based on historical data, preset thermal risk judgment criteria, trigger the thermal runaway early warning mechanism, and start a graded thermal protection strategy. S4: By acquiring the electromagnetic interference signal from the ozone power supply output side in real time, setting the electromagnetic interference safety threshold, and triggering the interference suppression response based on the electromagnetic interference signal, the power supply operating frequency is first adjusted using a filtering adjustment mechanism, generating an electromagnetic interference exceeding the standard fault report and simultaneously recording the relevant exceeding data.

[0027] Specifically, the construction process of the insulation safety assessment model is as follows: First, obtain the insulation operation status records and fault cases throughout the entire life cycle of the power supply. Combine the insulation failure mechanism analysis to screen the characteristic indicators that affect safety. Calculate the correlation weight of the characteristic indicators with insulation safety through statistical modeling. Then, build the core operation framework of the model using a weighted comprehensive evaluation mechanism and simultaneously integrate the insulation characteristic difference data under different operating conditions for dynamic calibration.

[0028] Specifically, the multi-level safety threshold system is generated as follows: statistically clustering historical insulation fault data, screening influencing factors and calculating weights, dynamically correcting the threshold boundary by simulating the insulation failure evolution process, and continuously obtaining fault data to iteratively optimize the threshold range.

[0029] Specifically, the specific generation process of the insulation fault location report is as follows: real-time acquisition of insulation anomaly parameters, feature extraction, construction of a feature vector set, similarity calculation with the multi-dimensional fault feature matrix in the preset fault template library, simultaneous application of time series analysis mechanism to dynamically model the fault evolution process, and based on the coupling analysis of similarity calculation results and trend prediction model, the fault location is locked, and finally, an insulation fault location research report is generated.

[0030] In this embodiment, a certain type of industrial-grade high-frequency ozone power supply system is used as the application object. The specific process, addressing the insulation safety protection requirements of its main circuit, is as follows: Real-time acquisition of insulation status parameters: Insulation monitoring sensors are deployed on the bus side, power module interface side, and load connection terminal of the ozone power supply main circuit to collect three types of core insulation status parameters in real time: including insulation resistance (denoted as R). a Rᵦ and Rᵧ (corresponding to different monitoring points), leakage current (denoted as I1 and I2), and partial discharge amplitude (denoted as P). x The acquisition period is set to T, continuously acquiring parameter data and transmitting it to the system control unit to form a real-time parameter sequence.

[0031] Insulation safety assessment model construction: First, the insulation operation status records of this type of ozone power source throughout its entire life cycle (covering four stages: normal operation, slight degradation, moderate degradation, and severe failure) and historical failure cases (the cumulative number of cases is denoted as N) are collected. Combined with the insulation failure mechanism analysis, insulation aging, humidity influence, mechanical loss, and electrical stress impact are identified as the core failure causes. From these, key characteristic indicators are selected: insulation resistance attenuation rate (denoted as Kᵣ), leakage current growth rate (denoted as Kᵢ), and partial discharge frequency (denoted as F). p ).

[0032] Feature weight calculation and model framework construction: Using the analytic hierarchy process (AHP) in statistical modeling, the correlation weights of the above feature indicators are calculated, resulting in weights w1, w2, and w3 for each indicator (satisfying w1 + w2 + w3 = 1). A core computational framework for the model is built based on a weighted comprehensive evaluation mechanism, with the evaluation value calculated using the formula: S = w1 × Kᵣ + w2 × Kᵢ + w3 × F. p , where S is the comprehensive evaluation value of insulation safety.

[0033] Model dynamic calibration: Insulation characteristic difference data of the ozone power source under different operating conditions (including rated power condition Q1, low power condition Q2, and start-stop transition condition Q3) were collected. A condition correction coefficient matrix was established (denoted as C=[c1,c2,c3], corresponding to the three operating conditions). The correction coefficients were then incorporated into the evaluation model. The final calibrated evaluation value is: S'=S×C k (C)k (This is the correction coefficient corresponding to the current working condition), enabling the model to dynamically adapt to different working conditions.

[0034] Generation of a multi-level security threshold system: K-means clustering was performed on historical insulation fault data (sample size denoted as M) to divide the faults into three clusters: mild risk, moderate risk, and severe risk. Three influencing factors were identified: critical insulation resistance, peak leakage current, and partial discharge threshold. Their weights, w, were calculated using the entropy weight method. a wᵦ w c .

[0035] Dynamic threshold boundary correction: An insulation failure evolution model is built using simulation software to simulate the entire process of insulation state from normal to failure under different influencing factors, dynamically correcting the threshold boundaries for each risk level. This ultimately generates a three-level safety threshold system. Safe interval: S'∈[A,B] (no protective action); Warning interval: S'∈[B,C] (Level 1 protection action); Fault range: S'∈[C,D] (Level 2 protection action); Threshold interval iterative optimization: During system operation, new fault data is continuously acquired (optimization is triggered when the cumulative number of new samples is denoted as L). The threshold intervals [A,B], [B,C], and [C,D] are iteratively updated through a feedback adjustment mechanism to ensure the accuracy and timeliness of the threshold system.

[0036] The control unit calculates the comprehensive insulation safety calibration value S' in real time and determines its threshold range: When S'∈[A,B], maintain the normal operation of the power supply and continuously monitor parameter changes; When S'∈[B,C], a first-level protection action is triggered: a local audible and visual warning signal is issued, the power output power is simultaneously reduced to η times the rated power, and the parameter acquisition cycle is shortened to T / 2; When S'∈[C,D], the secondary protection action is triggered: the main circuit isolation switch is immediately cut off, power restart is prohibited, and the fault location process is initiated.

[0037] Insulation fault location report generation: After the secondary protection action is triggered, the insulation abnormality parameters (including R) of the current time and the t consecutive acquisition cycles before the fault are extracted. a 、Rᵦ、Rᵧ、I1、I2、P x The feature vector set V=[v1,v2,v3,v4,v5,v6] is constructed through feature engineering (each element corresponds to a standardized feature value with different parameters).

[0038] The feature vector V is compared with the multidimensional fault feature matrix (denoted as M=[M1,M2,M3], corresponding to bus-side insulation fault, module interface-side insulation fault, and load-side insulation fault, respectively) in the preset fault template library using cosine similarity calculation to obtain similarity values ​​S1, S2, and S3. Simultaneously, a time-series analysis mechanism is used to dynamically model the fault evolution process (from parameter anomaly to protection triggering), generating a trend curve Tᵣ.

[0039] Based on the coupled analysis of similarity calculation results (taking the fault type corresponding to the maximum value) and trend prediction models, the location of the fault is pinpointed (such as insulation breakdown on the busbar side, insulation aging on the module interface side, etc.). The fault location information, feature vector data, similarity calculation results, trend curves, and current operating conditions are integrated to generate a standardized insulation fault location report, which is simultaneously uploaded to the system monitoring platform, and a retrieval index is established (index identifier ID=α) to facilitate subsequent operation and maintenance queries and traceability.

[0040] Specifically, the backflow determination conditions include: based on the energy backflow signal, firstly, synchronously acquire the instantaneous values ​​of load-side voltage and current, and through a data fusion mechanism, calculate the instantaneous power direction and amplitude change. During the backflow determination stage, calculate the deviation rate between the real-time calculation results and the preset benchmark value, and simultaneously monitor the duration of the deviation.

[0041] Specifically, the bidirectional energy feedback and absorption mechanism includes: firstly, using a real-time grid parameter monitoring algorithm to quantitatively assess the grid's dynamic acceptance capacity; based on the assessment results, triggering the bidirectional energy feedback and absorption mechanism to safely inject system backflow energy into the grid; and activating an energy caching strategy to temporarily store excess backflow energy in an energy storage unit while continuously monitoring grid parameters.

[0042] In this embodiment, an industrial-grade high-frequency ozone power supply (compatible with dielectric barrier discharge type ozone generator load) is used as the application object, and the specific process is as follows: Real-time acquisition of reverse current signal and load-side parameters: By deploying high-frequency voltage sensors, current sensors, and a dedicated energy backflow monitoring module at the output end of the ozone power supply load side, key data can be collected in real time. Energy backflow signal (denoted as Pᵣ): Characterizes the strength of the energy signal transmitted from the load side to the power supply side in reverse. It has no physical dimension and is only used to determine the backflow trend. Instantaneous value of load-side voltage (denoted as U) l ): This refers to the real-time instantaneous voltage value at the load input terminal of the ozone generator, measured in volts (V), which is the basic data for calculating power. Instantaneous value of load-side current (denoted as I) l): This refers to the instantaneous value of the real-time current flowing through the ozone generator load, measured in amperes (A). It is used in conjunction with voltage to perform power-related calculations. The parameter acquisition period is set to T2 (a fixed time interval to ensure real-time signal transmission). The acquired data is synchronously uploaded to the system control unit via a high-speed transmission channel.

[0043] Execution of reflux determination conditions: After receiving the data, the control unit executes the following steps according to the preset backflow determination logic: Data fusion and core parameter calculation: A Kalman filter data fusion mechanism is used to process synchronously acquired U... l I l Noise filtering and data calibration are performed, and two core metrics are calculated based on the calibrated data: Instantaneous power direction (denoted as P) n The positive direction (energy flowing from the power source to the load) is marked with "+", and the negative direction (energy flowing back from the load to the power source) is marked with "-". This is the core basis for determining the direction of energy transmission. Amplitude change (denoted as ΔP): the difference between the current instantaneous power and the power in the previous acquisition cycle (ΔP = |P_0.05) nk -P nk-1 |, k is the current acquisition time, k-1 is the previous acquisition time), used to characterize the power fluctuation amplitude and assist in judging the intensity of the backflow impact.

[0044] Deviation rate calculation and duration monitoring: The preset reference values ​​include the power amplitude reference value P0 (the standard value of instantaneous power normally absorbed by the load side under rated operating conditions) and the amplitude change reference value ΔP0 (the upper limit of the allowable normal power fluctuation). Both are reference thresholds preset based on the rated parameters of the equipment. Deviation rate calculation: Calculate the deviation rate δ between the real-time instantaneous power amplitude and P0 (δ=|(P0) / P0) n The deviation rate δ' (δ'=|(ΔP-ΔP0) / ΔP0|×100%) is calculated simultaneously with the amplitude-P0(ΔP-ΔP0) / P0|×100%). The deviation rate is used to quantify the degree of deviation between the real-time parameter and the standard value. Continuous duration monitoring: Activate the timing module and record "δ≥δ t (Preset deviation rate threshold) and δ'≥δ' t The continuous duration t of "(preset amplitude change deviation rate threshold)" s t0 is a preset duration threshold used to avoid misjudgments caused by instantaneous fluctuations. When "P" is satisfied simultaneously... n For '-' (energy reverse), δ≥δ t , δ'≥δ' t t sWhen the value is greater than or equal to t0, it is determined to be an effective energy backflow, triggering the subsequent protection mechanism.

[0045] The bidirectional energy feedback and absorption mechanism is activated: Quantitative assessment of the grid's dynamic acceptance capacity: A real-time power grid parameter monitoring algorithm is activated to synchronously collect key power grid parameters: power grid voltage U9, power grid frequency f9, and power grid phase φ9 (all three are core parameters reflecting the real-time operating status of the power grid). Based on these parameters, the dynamic acceptance capacity assessment value K of the power grid is calculated (the value range is 0≤K≤1, and the closer K is to 1, the more reverse energy the power grid can accept). The assessment logic is K=α×(U9 stability coefficient)+β×(f9 stability coefficient)+γ×(phase matching coefficient), where α, β, and γ are weighting coefficients (satisfying α+β+γ=1), used to balance the impact of different power grid parameters on the acceptance capacity.

[0046] Energy feedback and cache execution: Energy feedback: when K≥K t (Preset grid acceptance threshold, such as K) t When the value is 0.7 (used to determine whether the power grid has the conditions for energy acceptance), the bidirectional energy feedback mechanism is triggered. The bidirectional converter converts the reverse current energy on the load side into electrical energy that matches the voltage, frequency, and phase of the power grid, and injects it into the power grid safely to avoid energy waste. Energy caching: when K <K t When the grid's capacity is insufficient or the reverse flow energy exceeds the grid's real-time capacity, an energy caching strategy is activated to temporarily store the excess reverse flow energy in energy storage unit E. s (such as supercapacitor modules, E) max The rated capacity of the energy storage unit (i.e., the upper limit of the buffered energy) is used to prevent energy surges from damaging the equipment. Continuous monitoring: During the energy feedback and buffering process, U9, f9, φ9, and E are continuously monitored every T2 / 2 (shortening the acquisition cycle to improve monitoring sensitivity). s The real-time energy storage value is used to dynamically adjust the feedback power and buffer ratio to ensure the process is safe and stable.

[0047] Power control and main circuit disconnection: Power regulation: After determining that there is effective back current, the power supply output power is first reduced to β times the rated power (β<1, such as β=0.5, where β is the power reduction factor used to reduce energy conflict between the power supply side and the load side) to alleviate the back current intensity. Status monitoring: Continuously monitor the backflow signal Pᵣ and the instantaneous power direction P. n If the backflow condition is not relieved within t1 time after power adjustment (and the effective backflow judgment condition is still met), then the main circuit cut-off operation is executed. Switch disconnection and restart limitation: Control main circuit isolation switch S m Disconnect the power supply from the load and the power grid; at the same time, send a lock command to the power control chip to prevent the power supply from restarting until it is manually unlocked and the fault is investigated, so as to avoid repeated restarts that may cause secondary damage if the fault is not eliminated.

[0048] Reverse flow related operational data records: The system automatically records all relevant operational data for the entire reflux process, including the reflux determination time, reflux start and end times, and real-time U-turns. l / I l Sequence, P n / ΔP calculation results, δ / δ' deviation rate data s The data includes duration, grid parameters U9 / f9 / φ9, K-value variation curve, energy feedback / buffer quantity, power regulation trajectory, and switch disconnection time. All data is stored in categories of "date-operating condition-reverse current level," with a retrieval index ID=θ, and is synchronously uploaded to the system monitoring platform for subsequent fault tracing and protection strategy optimization.

[0049] Brief explanation of core parameters: The energy backflow signal Pᵣ is used to determine the trend of energy reverse transmission; U l I l It is the foundation of power calculation and directly reflects the electrical state of the load; P n Determining the direction of energy transfer is the core of backflow determination; ΔP characterizes power fluctuations and helps determine the degree of backflow impact; P0 and ΔP0 serve as reference benchmarks for normal operating conditions, while δ and δ' quantify the degree of parameter deviation; t s t0 avoids misjudgment and ensures accuracy; U9, f9, and φ9 are used to assess grid acceptance capacity; the K value determines the energy processing method; K t E represents the critical value of the grid's capacity to accept power. s E max Responsible for temporarily storing excess energy; β is the power reduction regulation ratio, S m The circuit cut-off actuator, with all parameters working together, achieves accurate determination and safety protection against energy backflow.

[0050] Specifically, the thermal runaway prediction model is constructed as follows: historical temperature time series data and thermal failure cases of the ozone power module are obtained, key temperature characteristic parameters of thermal runaway are screened by combining heat conduction mechanism analysis, the correlation strength between key temperature characteristic parameters of thermal runaway and thermal runaway risk is calculated by statistical modeling, and then the core computing framework is built by using a trend extrapolation mechanism, and thermal characteristic difference data under different operating conditions are incorporated for calibration.

[0051] Specifically, the thermal risk determination criteria include: combining the thermal conduction characteristics of the power module with the thermal runaway evolution law, integrating real-time temperature values, temperature rise rate, and temperature fluctuation amplitude as determination indicators, assigning judgment weights to the indicators according to the degree of thermal failure impact, performing thermal risk status determination through a multi-indicator linkage verification mechanism, and continuously iterating and correcting indicator weights and judgment thresholds based on newly acquired thermal failure cases and temperature monitoring data.

[0052] Specifically, the thermal runaway early warning mechanism is executed as follows: first, the early warning information is encapsulated in a structured manner, and the local warning logic and remote data transmission process are triggered simultaneously. At the same time, an early warning information traceability index is established, which is associated with the temperature parameters, the thermal risk level assessment results, and historical cases.

[0053] Specifically, the graded thermal protection strategy includes: simultaneously performing heat dissipation regulation and gradient power reduction operations, acquiring temperature data in real time to check the cooling effect, executing a power supply cut-off process based on the cooling effect, and gradually dissipating residual energy by utilizing energy release before the cut-off.

[0054] This embodiment takes the core power module (IGBT power module) of an industrial-grade high-frequency ozone power supply as the protection object, and the specific process is as follows: Real-time acquisition of power module temperature parameters: By deploying high-precision temperature sensors in the core chip area of ​​the power module, the heat dissipation surface of the casing, and the substrate interface, three key temperature parameters are collected in real time: Core module temperature (denoted as T_c): directly reflects the real-time heating status of power devices (such as IGBT chips), and is expressed in degrees Celsius (°C). It is a core indicator for thermal risk assessment. Module housing temperature (denoted as T_h): Characterizes the heat dissipation efficiency of the module housing, in °C, and helps to determine the working status of the heat dissipation system; Substrate temperature (denoted as T_b): reflects the change in thermal resistance of the heat conduction path, measured in °C, and is used to predict the risk of heat conduction failure. The parameter acquisition period is set to T3 (a fixed time interval to ensure the real-time nature of temperature data). The acquired data is uploaded to the system control unit through an isolated transmission channel to form a continuous temperature time sequence.

[0055] Construction of thermal runaway prediction model: Historical data collection and feature parameter selection: We collected historical temperature time-series data (cumulative sample duration denoted as T_total) and thermal failure cases (cumulative number of cases denoted as N_f) for the entire lifecycle of this type of power module. Combined with heat conduction mechanism analysis (covering the characteristics of three heat dissipation paths: conduction, convection, and radiation), we screened out the key temperature characteristic parameters for thermal runaway: Core temperature peak (denoted as T_cm): The maximum instantaneous value of T_c during module operation; Temperature rise rate (denoted as v_T): The increase in T_c per unit time (v_T=ΔT_c / Δt, where Δt is the difference between adjacent sampling periods); Temperature fluctuation range (denoted as A_T): The difference between the maximum and minimum values ​​of T_c within a set time window.

[0056] Correlation strength calculation and model framework construction: The Pearson correlation coefficient method in statistical modeling was used to calculate the correlation strength between the three types of characteristic parameters and the thermal runaway risk. The correlation coefficients were r1 (T_cm and thermal runaway risk), r2 (v_T and thermal runaway risk), and r3 (A_T and thermal runaway risk). Based on the correlation coefficients, weights were assigned to each characteristic parameter: w_a=r1 / (r1+r2+r3), w_b=r2 / (r1+r2+r3), w_c=r3 / (r1+r2+r3) (satisfying w_a+w_b+w_c=1). A trend extrapolation mechanism (integrating linear regression and time series prediction algorithms) is used to build the core computational framework of the model. The formula for calculating the thermal runaway risk assessment value R is: R=w_a×(T_cm / T_cmax)+w_b×(v_T / v_Tmax)+w_c×(A_T / A_Tmax), where T_cmax, v_Tmax, and A_Tmax are the safety limit values ​​of the three types of parameters, respectively.

[0057] Model dynamic calibration: Data on the thermal characteristics of the module under different operating conditions (rated power condition Q1, overload condition Q2, and light load condition Q3) are collected. A condition correction coefficient matrix M=[m1,m2,m3] (corresponding to the three types of operating conditions) is established. The correction coefficients are integrated into the evaluation model, and the calibrated risk assessment value R'=R×M_k (M_k is the correction coefficient corresponding to the current operating condition) is obtained to ensure the accuracy of the model's evaluation under different operating conditions.

[0058] Thermal risk assessment criteria are applied as follows: Thermal conduction characteristics of power modules: This refers to the heat transfer pattern inside ozone power modules (such as IGBT modules). Heat is conducted from the core heat-generating area (chip T_c) through the substrate (T_b) to the outer shell (T_h), and then dissipated through the heat dissipation system (air cooling or water cooling). The transfer efficiency is affected by the thermal resistance of the material, the contact gap, and the heat dissipation method, which directly determines the distribution and trend of temperature parameters. It is the basis for screening monitoring parameters such as T_c, T_h, and T_b.

[0059] Thermal runaway evolution law: It refers to the progressive process of the power module from normal heating to thermal runaway, following the evolution logic of "normal heat dissipation → heat accumulation (T_c continuously increases, v_T accelerates) → thermal resistance mutation (temperature fluctuation A_T increases) → thermal runaway (exceeding T_cmax)". The changes in temperature values, temperature rise rates, and fluctuation amplitudes have obvious time sequences, which are the core basis for setting the verification priority of "temperature value → temperature rise rate → fluctuation amplitude" and the hierarchical protection strategy.

[0060] Preset the threshold system for three types of judgment indicators: Real-time temperature value thresholds: T1 (warning threshold), T2 (danger threshold), T1 < T2 < T_cmax; Temperature rise rate thresholds: v1 (warning threshold), v2 (danger threshold), v1 < v2 < v_Tmax; Temperature fluctuation amplitude thresholds: A1 (warning threshold), A2 (danger threshold), A1 < A2 < A_Tmax. Perform linkage verification according to the priority of "temperature value → temperature rise rate → fluctuation amplitude". First, judge whether T_c exceeds the threshold, then verify v_T, and finally evaluate A_T. The verification results of each index are weighted and integrated according to the weights w_a, w_b, w_c to obtain the final judgment result.

[0061] Iterative correction of judgment thresholds: During the system operation process, every time N_f' new thermal fault cases are accumulated or the continuous operation duration reaches T_update, the criterion optimization process is automatically triggered: Based on the newly added temperature monitoring data and fault cases, recalculate the correlation strength of characteristic parameters, and correct the weights w_a, w_b, w_c and the thresholds of each index (T1, T2, v1, v2, A1, A2) to ensure that the judgment criterion is dynamically adapted to the module aging state and actual operating conditions.

[0062] Trigger of thermal runaway warning mechanism: Structured encapsulation of warning information: When the judgment result meets the warning conditions (R' ∈ [R1, R2), R1 is the warning critical value, and R2 is the danger critical value), the warning information is structurally encapsulated, including core information such as the warning trigger time, current temperature parameters (T_c, T_h, T_b), risk assessment value R', thermal risk level (primary warning / secondary warning), and current operating conditions.

[0063] Warning and data transmission: Simultaneously trigger the dual-channel warning logic: Locally trigger the audible and visual warning (the flashing frequency of the indicator light L is f1, and the buzzer B emits intermittent alarms); remotely transmit the warning information to the monitoring center platform through the industrial Ethernet, and the transmission delay ≤ t_delay.

[0064] Establishment of warning traceability index: Each early warning message is assigned a unique traceability index ID=λ, which is associated with and stored temperature time series data, risk level assessment results, and corresponding historical failure cases (based on R' similarity matching) to form a complete early warning traceability chain, facilitating subsequent analysis.

[0065] Implementation of graded thermal protection strategy: Level 1 protection (R'∈[R1,R2, mild thermal risk): Heat dissipation regulation: Activate the dynamic heat dissipation mechanism and set the adjustment coefficient of the heat dissipation system (such as forced air cooling + water cooling composite heat dissipation) to k1 (k1>1, indicating enhanced heat dissipation intensity) to increase heat dissipation power; Gradient power reduction: Controls the power supply output power to be reduced to η1 times the rated power (η1<1, such as η1=0.8), reducing the heat source of the module; Effect monitoring: The acquisition cycle is shortened every T3 / 2 interval to continuously monitor the trend of T_c change. If T_c drops below T1 within t4 time, the current protection status is maintained and the power is gradually restored; if the temperature does not drop, the protection is upgraded to level 2.

[0066] Secondary protection (R'∈[R2,R3, moderate thermal risk): Enhance heat dissipation: Increase the heat dissipation adjustment coefficient to k2 (k2>k1) and activate the backup heat dissipation channel (such as the emergency water cooling circuit). Deep power reduction: further reduce the power output to η2 times the rated power (η2<η1, such as η2=0.5). Effect verification: Continuously monitor T_c. If T_c is still not lower than T1 within t5, trigger level 3 protection.

[0067] Level 3 protection (R'≥R3, severe thermal risk / critical thermal runaway): Power supply cut-off procedure: The main circuit power switch S_p is opened, cutting off the module power supply circuit; Residual energy discharge: Before disconnection, the energy release mechanism is activated to gradually discharge the residual energy inside the module and the circuit through the discharge resistor R_d. The discharge time is ≥t6 to ensure that the residual energy is ≤E_res (safe residual energy threshold). Status Lock: After disconnection, a lock command is sent to prevent power restart until the fault is manually checked and unlocked.

[0068] Specifically, the filtering adjustment mechanism is executed as follows: first, the suppression parameters are adjusted based on the electromagnetic interference exceeding the standard frequency band; the changes in the interference signal after the filtering effect are monitored in real time; the power supply operating frequency is dynamically optimized according to the suppression effect; and the output voltage and power are maintained in a stable state through collaborative control logic.

[0069] Specifically, the electromagnetic interference exceeding the standard fault report generation process is as follows: based on the actual suppression effect after filter adjustment and power supply operating frequency adjustment, integrate electromagnetic interference characteristic parameters and corresponding filter parameters and frequency adjustment trajectory, synchronously associate with the current operating condition information of the power supply, classify the recorded exceeding data according to the operating condition type to establish a retrieval index, and immediately synchronize it to the system monitoring terminal after generation.

[0070] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for protecting the safe operation of an ozone power supply system, characterized in that, include: S1: Real-time acquisition of insulation status parameters of the main circuit, construction of insulation safety assessment model, generation of multi-level safety threshold system, execution of graded protection actions according to the threshold range of the insulation status parameters, and generation of insulation fault location report; S2: Real-time acquisition of energy backflow signal from the ozone power supply load side, setting backflow judgment conditions, starting bidirectional energy feedback and absorption mechanism, continuously monitoring backflow status, controlling power output power, cutting off main circuit switch, prohibiting power restart and recording backflow-related operating data; S3: Obtain the temperature parameters of the ozone power module, build a thermal runaway prediction model based on historical data, preset thermal risk judgment criteria, trigger the thermal runaway early warning mechanism, and start a graded thermal protection strategy. S4: By acquiring the electromagnetic interference signal from the ozone power supply output side in real time, setting the electromagnetic interference safety threshold, and triggering the interference suppression response based on the electromagnetic interference signal, the power supply operating frequency is first adjusted using a filtering adjustment mechanism, generating an electromagnetic interference exceeding the standard fault report and simultaneously recording the relevant exceeding data.

2. The method according to claim 1, characterized in that, The specific construction process of the insulation safety assessment model is as follows: First, obtain the insulation operation status records and fault cases throughout the entire life cycle of the power supply. Combine the insulation failure mechanism analysis to screen the characteristic indicators that affect safety. Calculate the correlation weight of the characteristic indicators with insulation safety through statistical modeling. Then, build the core operation framework of the model using a weighted comprehensive evaluation mechanism. Simultaneously, incorporate the insulation characteristic difference data under different operating conditions for dynamic calibration.

3. The method according to claim 1, characterized in that, The specific generation process of the multi-level safety threshold system is as follows: statistically cluster historical insulation fault data, screen influencing factors and calculate weights, dynamically correct the threshold boundary by simulating the insulation failure evolution process, and continuously obtain fault data to iteratively optimize the threshold range.

4. The method according to claim 1, characterized in that, The specific generation process of the insulation fault location report is as follows: real-time acquisition of insulation anomaly parameters, feature extraction, construction of a feature vector set, similarity calculation with the multi-dimensional fault feature matrix in the preset fault template library, simultaneous application of time series analysis mechanism to dynamically model the fault evolution process, and coupling analysis based on similarity calculation results and trend prediction model to locate the fault location and finally generate an insulation fault location research report.

5. The method according to claim 1, characterized in that, The backflow determination conditions specifically include: based on the energy backflow signal, firstly, synchronously acquire the instantaneous values ​​of load-side voltage and current, and through a data fusion mechanism, calculate the instantaneous power direction and amplitude change. During the backflow determination stage, calculate the deviation rate between the real-time calculation results and the preset benchmark value, and simultaneously monitor the duration of the deviation.

6. The method according to claim 1, characterized in that, The bidirectional energy feedback and absorption mechanism specifically includes: firstly, using a real-time grid parameter monitoring algorithm to quantitatively assess the grid's dynamic acceptance capacity; based on the assessment results, triggering the bidirectional energy feedback and absorption mechanism to safely inject the system's reverse flow energy into the grid; and activating an energy caching strategy to temporarily store excess reverse flow energy in an energy storage unit while continuously monitoring grid parameters.

7. The method according to claim 1, characterized in that, The specific construction process of the thermal runaway prediction model is as follows: obtain historical temperature time series data and thermal failure cases of the ozone power module, analyze and screen key temperature characteristic parameters of thermal runaway by combining heat conduction mechanism, calculate the correlation strength between key temperature characteristic parameters of thermal runaway and thermal runaway risk through statistical modeling, and then build the core computing framework by using trend extrapolation mechanism, and incorporate thermal characteristic difference data under different operating conditions for calibration.

8. The method according to claim 1, characterized in that, The thermal risk assessment criteria specifically include: combining the thermal conduction characteristics of the power module with the evolution law of thermal runaway, integrating real-time temperature values, temperature rise rate, and temperature fluctuation amplitude assessment indicators, assigning assessment weights to the indicators according to the degree of impact of thermal failure, performing thermal risk status assessment through a multi-indicator linkage verification mechanism during the process, and continuously iterating and correcting the indicator weights and assessment thresholds based on newly acquired thermal failure cases and temperature monitoring data.

9. The method according to claim 1, characterized in that, The thermal runaway early warning mechanism specifically includes: firstly, the early warning information is structured and encapsulated, and the local warning logic and remote data transmission process are triggered simultaneously. At the same time, an early warning information traceability index is established, which is associated with the temperature parameters, the thermal risk level assessment results, and historical cases.

10. The method according to claim 1, characterized in that, The graded thermal protection strategy specifically includes: simultaneously performing heat dissipation regulation and gradient power reduction operations, acquiring temperature data in real time to check the cooling effect, and based on the cooling effect, executing a power supply cut-off process, and gradually dissipating residual energy by utilizing energy release before the cut-off.

11. The method according to claim 1, characterized in that, The filtering and adjustment mechanism specifically includes: first, adjusting the suppression parameters based on the electromagnetic interference exceeding the standard frequency band; monitoring the changes in the interference signal after filtering in real time; dynamically optimizing the power supply operating frequency according to the suppression effect; and maintaining the stable state of output voltage and power through collaborative control logic.

12. The method according to claim 1, characterized in that, The electromagnetic interference exceeding the standard fault report generation process is as follows: based on the actual suppression effect after filter adjustment and power supply operating frequency adjustment, the electromagnetic interference characteristic parameters and corresponding filter parameters and frequency adjustment trajectory are integrated, the current operating condition information of the power supply is synchronously associated, the recorded exceeding data are classified according to the operating condition type to establish a retrieval index, and after generation, it is immediately synchronized to the system monitoring terminal.

Images