Safety protection method, system and equipment for audio equipment and medium

Abstract

The invention relates to a safety protection method and system for audio equipment, equipment and a medium. The method comprises the following steps: firstly, constructing and calibrating a static equipotential basic network comprising a variable impedance unit to obtain a reference impedance characteristic of the static equipotential basic network, and then synchronously collecting potential data of each key node in the network and final audio output signal data in real time; based on the data and a reference model, constructing a multi-objective optimization function taking minimization of dangerous power frequency potential difference (safe potential difference) and output noise as core objectives, and dynamically generating and executing an optimal control instruction for each variable impedance unit by solving the function; therefore, the impedance distribution of the whole equipotential network in a broadband can be actively and accurately adjusted, and finally, the dangerous potential difference between touchable components in the cabinet can be continuously suppressed in a complex operation environment so as to guarantee the personal safety. And meanwhile, audio noise introduced by a grounding loop and interference is effectively reduced, so that the signal quality is improved.

Description

Technical Field

[0001] This invention belongs to the field of artificial intelligence and automation control, and in particular relates to a method, system, device and medium for the security protection of audio devices. Background Technology

[0002] In professional audio systems, such as those used in large concerts, theaters, or stadiums, multiple high-power amplifiers, digital processors, and other equipment are typically densely packed within metal cabinets. These devices are powered by the public power grid, and their metal casings, signal grounds, and the cabinet itself must all be grounded to meet electrical safety regulations and ensure stable system operation.

[0003] However, this complex interconnected system faces a fundamental contradiction in practice: on the one hand, to prevent the risk of electric shock caused by equipment leakage, lightning induction, or power grid failure, all accessible metal components inside the cabinet must be connected at low impedance equipotential bonding using robust conductors to eliminate dangerous potential differences between them; on the other hand, this low-impedance parallel grounding network is prone to forming a massive "ground loop" with the audio signal connection lines between devices. Alternating magnetic fields in space (such as power transformers and lightning electromagnetic pulses) induce currents in the ground loop. These currents are converted into noise voltages on the parasitic impedance of the grounding conductor and are introduced into the sensitive audio signal path, resulting in persistent AC hum, buzzing, or radio frequency interference in the output audio, severely degrading sound quality. Traditionally, audio engineers have tried to suppress noise using techniques such as "star-topology single-point grounding" of audio signals, isolation transformers, or balanced transmission. However, these methods are often difficult to implement thoroughly in complex multi-device cabinets and may weaken the discharge path for high-frequency interference, or even conflict with the requirements of safe grounding in extreme cases.

[0004] Therefore, existing technologies are essentially static or compromise solutions: fixed-connection equipotential networks ensure basic safety but sacrifice audio purity; while grounding optimization focused on sound quality may introduce potential safety hazards and cannot cope with dynamic changes in the power grid environment and random start-ups and shutdowns of equipment in real time. The core problem lies in the lack of an intelligent method that can sense the internal potential distribution and noise status of the system in real time and dynamically adjust the impedance characteristics of the grounding network to simultaneously and proactively maintain optimal safety protection and the highest signal integrity. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, system, device, and medium for the security protection of audio devices to address the aforementioned technical problems.

[0006] Firstly, this application provides a security protection method for an audio device, including:

[0007] S1. Establish a static equipotential foundation network for the audio equipment, perform broadband impedance parameter calibration on the static equipotential foundation network, and obtain the network reference impedance matrix; wherein, the main grounding reference point, the audio signal star grounding bus, and the preset positions of the potential connection architecture in the static equipotential foundation network are equipped with variable impedance units.

[0008] S2. Synchronously sample the ground potential of multiple monitoring nodes set on the static equipotential foundation network to obtain potential data; synchronously collect the audio output signal of the power amplifier of the audio equipment to obtain audio data.

[0009] S3. Based on the network reference impedance matrix, potential data, and audio data, a multi-objective optimization function is constructed with the goal of minimizing the safety potential difference and output audio noise. By solving the multi-objective optimization function, an optimal control instruction set is generated for dynamically adjusting the impedance characteristics of each variable impedance unit. Here, the safety potential difference represents the power frequency potential difference between any two accessible metal parts in the audio device that could cause the risk of electric shock.

[0010] S4. Drive each variable impedance unit to adjust its impedance value in a wide frequency range according to the optimal control instruction set;

[0011] S5. Continuously monitor the ground potential and audio output signal noise of the monitoring node. When an abnormal potential or a new noise event is detected, repeat S2 to S4.

[0012] Secondly, this application also provides a security protection system for an audio device, used to implement the method described in the first aspect, the system comprising:

[0013] The network reference construction module is used to establish a static equipotential base network for audio equipment, perform broadband impedance parameter calibration on the static equipotential base network, and obtain the network reference impedance matrix. Variable impedance units are set at preset positions of the main grounding reference point, the audio signal star grounding bus, and the potential connection architecture in the static equipotential base network.

[0014] The real-time monitoring and acquisition module is used to synchronously sample the ground potential of multiple monitoring nodes set on the static equipotential foundation network to obtain potential data; and to synchronously acquire the audio output signal of the power amplifier of the audio equipment to obtain audio data.

[0015] The multi-objective optimization decision module is used to construct a multi-objective optimization function based on the network reference impedance matrix, potential data, and audio data, with the optimization objectives being to minimize the safety potential difference and output audio noise. By solving the multi-objective optimization function, an optimal set of control instructions is generated to dynamically adjust the impedance characteristics of each variable impedance unit. Here, the safety potential difference represents the power frequency potential difference between any two accessible metal parts in the audio device that could cause the risk of electric shock.

[0016] The impedance dynamic adjustment module is used to drive each variable impedance unit to adjust its impedance value over a wide frequency band according to the optimal control instruction set.

[0017] The adaptive feedback monitoring module is used to continuously monitor the ground potential and noise of the audio output signal of the monitoring node. When an abnormal potential or a new noise event is detected, the real-time monitoring and acquisition module, the multi-objective optimization decision-making module, and the impedance dynamic adjustment module are activated to re-execute the corresponding operations.

[0018] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a security protection method for an audio device as described in the first aspect.

[0019] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a security protection method for an audio device as described in the first aspect.

[0020] The aforementioned method, system, device, and medium for the safety protection of audio equipment first constructs and calibrates a static equipotential base network containing variable impedance units to obtain its reference impedance characteristics. Then, it synchronously and in real-time collects the potential data of each key node in the network and the final audio output signal data. Based on these data and the reference model, it constructs a multi-objective optimization function with the core objectives of minimizing the dangerous power frequency potential difference (safe potential difference) and output noise. By solving this function, it dynamically generates and executes the optimal control commands for each variable impedance unit, thereby actively and accurately adjusting the impedance distribution of the entire equipotential network in a wide frequency band. Ultimately, it achieves the dual technical effects of continuously suppressing the dangerous potential difference between accessible components in the cabinet to ensure personal safety in complex operating environments, while effectively reducing audio noise introduced by grounding loops and interference to improve signal quality. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating a security protection method for an audio device provided by the present invention;

[0023] Figure 2 This is a schematic diagram illustrating the process of constructing and solving a multi-objective optimization function in one optional embodiment of the present invention;

[0024] Figure 3 This is a schematic diagram of the structure of a security protection system for an audio device provided by the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0026] refer to Figure 1 The document presents a flowchart illustrating a security protection method for an audio device provided in this application, which includes the following steps:

[0027] S1. Establish a static equipotential foundation network for the audio equipment, perform broadband impedance parameter calibration on the static equipotential foundation network, and obtain the network reference impedance matrix; wherein, the main grounding reference point, the audio signal star grounding bus, and the preset positions of the potential connection architecture in the static equipotential foundation network are equipped with variable impedance units.

[0028] Specifically, the construction of the static equipotential bonding network must comply with the IEC 61140 electrical safety standard and the audio system grounding specifications. Using a metal cabinet as the carrier, a high-conductivity copper busbar can be used as the main grounding trunk line, with a main grounding reference point located at the center of the cabinet's bottom. This reference point is connected to the building's grounding grid via a grounding wire, and the grounding wire impedance must be controlled within a range sufficient for rapid discharge of leakage current and lightning-induced current. The audio signal star grounding busbar can be made of high-conductivity oxygen-free copper and installed in the middle layer of the cabinet's back panel. Terminals are pre-installed on the busbar according to the number of devices. The signal ground leads of each audio device are individually connected to this busbar to avoid mutual interference between signal grounds, forming a star-shaped single-point grounding basic architecture.

[0029] The potential connection architecture reliably connects the cabinet frame, the metal casings of each device, the power distribution unit casing, and the main grounding reference point using copper braided tape. The cross-sectional area of ​​the braided tape must meet the low contact impedance requirement, and the connection points are secured with bolts and tin-plated to reduce contact impedance. The variable impedance unit uses a wideband programmable impedance module with a built-in MOSFET array, high-frequency choke, and precision resistor network, providing continuous impedance adjustment capability from power frequency to radio frequency. Its installation position is precisely set between the main grounding reference point and the star grounding bus, at each branch node of the bus, and on the connection path between the cabinet frame and the main grounding trunk line. Each variable impedance unit integrates an isolated drive interface and a status feedback module, supporting remote control and real-time impedance monitoring.

[0030] The broadband impedance parameter calibration process is completed collaboratively using a vector network analyzer and a high-precision grounding resistance tester. Before calibration, all external interference sources must be disconnected, and the system must be placed in a shielded environment. First, the variable impedance unit is set to an initial pass-through state, and the inherent impedance values ​​between each monitoring node are measured to establish an initial impedance matrix. Then, the impedance values ​​of each variable impedance unit are adjusted point-by-point within the frequency range from power frequency to radio frequency, with reasonable increments. Simultaneously, impedance parameters between nodes at the corresponding frequencies are collected, including resistance and reactance components. A network reference impedance matrix containing the frequency-impedance mapping relationship is generated through matrix operations. The specific calculation of the network reference impedance matrix uses the complex impedance synthesis formula: ,in For frequency Next The monitoring node and the first Complex impedance between monitoring nodes; For frequency Next The monitoring node and the first The resistance components between each monitoring node are directly acquired by an impedance tester; For frequency Next The monitoring node and the first The reactance components between monitoring nodes are expressed by the formula Calculations show that For the frequency measured by the vector network analyzer Next The monitoring node and the first The impedance amplitude between each monitoring node. The matrix dimension corresponds to the number of monitoring nodes, and the element values ​​are stored in complex number form to accurately characterize the impedance characteristics of the static network at different frequencies. The calibration data is filtered and denoised before being stored in the database of the system control unit as a benchmark for subsequent optimization calculations.

[0031] S2. Synchronously sample the ground potential of multiple monitoring nodes set on the static equipotential foundation network to obtain potential data; synchronously collect the audio output signal of the power amplifier of the audio equipment to obtain audio data.

[0032] Specifically, the monitoring nodes must be deployed to cover key locations in the static equipotential bonding network, and the number of nodes must meet comprehensive monitoring requirements. These nodes should be located at the main grounding reference point, both ends of the star grounding busbar, the power amplifier housing, the digital processor signal ground interface, the top frame of the cabinet, the bottom frame of the cabinet, the grounding terminal of the power distribution unit, and the housing of the audio signal output interface. Each monitoring node should be equipped with a high-precision differential potential sensor. The sensor's accuracy, bandwidth range, and isolation voltage must meet the requirements to avoid introducing interference and ensure the accuracy of the monitoring data.

[0033] The ground potential synchronous sampling adopts a distributed synchronous acquisition architecture. Each sensor achieves sampling timing alignment through a synchronization clock module. The synchronization clock module is designed based on a combination of GPS timing and a local high-precision crystal oscillator, with synchronization errors controlled within a negligible range. The sampling rate and resolution must meet the requirements for accurate acquisition of potential signals. The acquired data is transmitted to the central control unit via a high-speed bus. During the sampling process, the potential data undergoes real-time preprocessing. An FIR low-pass filter is used to filter out high-frequency glitches, and an abnormal pulse signal is eliminated through a moving average algorithm to ensure the stability of the potential data.

[0034] Audio data acquisition is conducted on the audio output signal of the power amplifier. Acquisition points are selected at the positive, negative, and ground terminals of the power amplifier's balanced output interface. A professional audio acquisition card is used, with a sampling rate and bit depth sufficient to support full-band audio signal acquisition. A shielded balanced cable connects the acquisition card and the power amplifier, with the cable shield grounded at one end to a star grounding busbar to avoid introducing additional interference. Audio and electrical potential data are synchronized using a timestamp mechanism. Each sampled data frame includes a precise timestamp, node number, and device identifier to ensure consistency between the two types of data in the time dimension.

[0035] After data acquisition, preliminary analysis was performed on the audio data to extract total harmonic distortion (THD), noise level, and noise amplitude in specific frequency bands. These specific frequency bands included the AC hum band and the radio frequency interference band. The noise level was calculated using the RMS method, with the specific formula as follows: In the formula The effective value of the noise signal is obtained through the formula. calculate, The number of sampling points. For the first The noise signal amplitude at each sampling point; The reference audio signal amplitude is taken as the peak value of the rated output audio signal. The total harmonic distortion calculation formula is: ,in The amplitude of the fundamental signal. to These represent the amplitudes of the 2nd to nth harmonic signals. The extracted parameters and potential data are stored together in a circular buffer. The buffer capacity must meet the data cyclic coverage requirements, while also reserving a data export interface for later troubleshooting and algorithm optimization.

[0036] S3. Based on the network reference impedance matrix, potential data, and audio data, a multi-objective optimization function is constructed with the goal of minimizing the safety potential difference and output audio noise. By solving the multi-objective optimization function, an optimal control instruction set is generated for dynamically adjusting the impedance characteristics of each variable impedance unit. Here, the safety potential difference represents the power frequency potential difference between any two accessible metal parts in the audio device that could cause the risk of electric shock.

[0037] Specifically, the process begins with preprocessing and feature extraction based on the network reference impedance matrix, potential data, and audio data. The actual potential difference between each monitoring node is calculated using the potential data, and the power frequency potential difference between accessible metal parts is selected as a safety assessment indicator. A safety potential difference threshold is set according to the IEC 60950 standard; exceeding this threshold indicates a risk of electric shock. Regarding audio data, the time-domain audio signal is converted to a frequency-domain signal using a fast Fourier transform. The noise amplitude of the corresponding frequency band for AC hum and the radio frequency interference band is extracted, and the ratio of the noise level to the original audio signal level is calculated as an audio quality assessment indicator, aiming to control the noise level within a preset low range.

[0038] The multi-objective optimization function has two objectives: minimizing the exceedance of the safety potential difference and minimizing the audio noise level. The independent variables are the impedance values ​​of each variable impedance unit at different frequencies. The exceedance of the safety potential difference is calculated using a piecewise function, with the specific formula as follows: In the formula The actual power frequency potential difference between accessible metal components is expressed by the formula... calculate, It is the power frequency angular frequency. For the first The impedance of the grounding path at power frequency. This is the grounding current for this path. For the number of grounding paths, The safe potential difference threshold is used. The audio noise level optimization target adopts a normalized processing formula: ,in This is the theoretical minimum noise level. To allow the maximum noise level, the final multi-objective optimization function takes the form: , It is a collection of variable impedance units. Representing the The impedance value (in ohms) of each variable impedance element at the corresponding frequency. This represents the total number of variable impedance units. and For the weighting coefficients, satisfying ,and , , The weighting adjustment coefficient (ranging from 0.6 to 0.9) enables adaptive adjustment under operating conditions. Constraints include: the impedance value of the variable impedance unit at the power frequency must be sufficiently small to ensure unobstructed leakage current discharge; the impedance value at the radio frequency band can be dynamically adjusted, but must meet power capacity requirements to avoid damage from instantaneous high current; the safe potential difference between nodes must not exceed a set threshold, and the audio noise level must not exceed a preset standard.

[0039] In the process of constructing the optimization function, a weighted summation method is used to transform the bi-objective function into a single-objective function. The transformation formula is as follows: .in This is the transformed single-objective optimization function, and the meanings of each parameter are consistent with those in the multi-objective function. The weighting coefficients are dynamically adjusted according to real-time operating conditions, increasing when there are large fluctuations in the grid voltage. Prioritizing safety, increase the frequency when the audio signal is weak. The focus is on sound quality optimization. A non-dominated sorting genetic algorithm is used for the solution process. This algorithm is suitable for multi-constraint, multi-objective optimization problems and can generate Pareto optimal solution sets. In the algorithm initialization phase, reasonable population size, number of iterations, crossover probability, and mutation probability are set. The impedance values ​​in the network's baseline impedance matrix are used as the initial population. During the iteration process, the optimal individual is selected using a fitness function, the expression of which is: ,in The fitness value is the individual's fitness value, which ranges from (0,1). The closer the value is to 1, the better the individual's performance.

[0040] During the iteration process, each generation of individuals corresponds to a set of variable impedance unit impedance combinations. The safety potential difference and audio noise level under this combination are calculated through simulation and used as the individual's fitness value. The optimal individual is selected through non-dominated sorting and crowding calculation, iterating gradually until convergence. After convergence, the solution with the best overall performance is selected from the Pareto optimal solution set and transformed into a wideband impedance adjustment curve for each variable impedance unit, generating a control command set. The control command set includes a frequency-impedance mapping table for each variable impedance unit in the power frequency to radio frequency range, adjustment rate, and feedback verification parameters. The commands are transmitted in an encrypted format to ensure control security, while an emergency command channel is reserved to deal with unexpected operating conditions.

[0041] S4. Drive each variable impedance unit to adjust its impedance value within a wide frequency band according to the optimal control instruction set.

[0042] Specifically, the central control unit first parses the control command set, converting the frequency-impedance mapping table into segmented control signals for each variable impedance unit, and employing differentiated control strategies for the power frequency band, intermediate frequency band, and radio frequency band. For the power frequency band, low impedance characteristics are prioritized, with control signals primarily consisting of constant impedance values. A PWM pulse width modulation signal drives the MOSFET array within the variable impedance unit to conduct, adjusting the duty cycle to achieve precise impedance control. Simultaneously, a built-in resistor network assists in calibration, ensuring impedance errors are kept within acceptable limits.

[0043] The mid-frequency and radio frequency (RF) bands employ a dynamic tracking control mode based on the impedance adjustment curve in the control command. This involves real-time adjustment of the PWM signal frequency and duty cycle to respond to impedance differences caused by frequency variations. A high-frequency choke plays a crucial role in suppressing RF signal leakage, preventing interference from impedance adjustment to the audio signal. The variable impedance unit is driven by an isolated drive module. The drive voltage and isolation voltage must meet the requirements for isolating high-voltage interference, effectively isolating high-voltage interference from the grounding network and the control circuit, preventing damage to the control unit.

[0044] During the adjustment process, the impedance sensor built into each variable impedance unit collects the actual impedance value in real time. The sampling frequency must meet the closed-loop adjustment requirements. The collected data is transmitted to the central control unit through the feedback channel and compared with the target impedance value in the control command. A PID control algorithm is then used for closed-loop adjustment. The PID control algorithm uses a position-based calculation method, and the specific formula is as follows: In the formula For the first The control signal quantity output to the variable impedance unit per sampling period For the first Impedance deviation value per sampling period, For the first The target impedance value for each sampling period For the first The actual acquisition impedance value for each sampling period, The sampling period is The proportionality coefficient is obtained through the formula. calculate, To control the maximum output amplitude of the signal, For the maximum permissible impedance deviation, , The integral time constant is... , This is the differential time constant. The control signal is dynamically calculated using this formula to correct the duty cycle and frequency parameters of the PWM signal, thereby calibrating the impedance deviation and ensuring the stability and accuracy of the adjustment process.

[0045] Simultaneously, the operating status of the variable impedance unit is monitored in real time, including operating temperature, current value, and voltage value. When any parameter exceeds the safe range, the protection mechanism is automatically activated, pausing impedance adjustment and switching to bypass mode. The variable impedance unit is short-circuited through the built-in bypass switch to maintain the safe grounding state of the static equipotential network. Adjustment resumes after the fault is cleared. After impedance adjustment is completed, the central control unit verifies the final impedance value of each variable impedance unit and compares it with the target value of the control command. If the verification is successful, the current impedance parameter is recorded, and the system impedance matrix is ​​updated as the benchmark for the next round of monitoring and adjustment. If the verification fails, the adjustment process is repeated until the accuracy requirements are met. The entire adjustment process is completed within a preset time to avoid significant impact on audio output quality. At the same time, the adjustment rate must be controlled within a reasonable range to prevent potential surges caused by impedance changes.

[0046] S5. Continuously monitor the ground potential and audio output signal noise of the monitoring node. When an abnormal potential or a new noise event is detected, repeat S2 to S4.

[0047] Specifically, the continuous monitoring adopts a hierarchical monitoring architecture. The first level of monitoring is real-time sampling monitoring, with the monitoring frequency consistent with step S2. The central control unit analyzes the sampled data in real time, calculates the safety potential difference, audio noise level and rate of change between each node, and generates monitoring logs. The logs contain information such as timestamps, monitoring parameter values, and device status. They are stored in a circular manner to retain monitoring data for a preset duration and support querying and exporting by the host computer.

[0048] Level 2 monitoring is for anomaly threshold monitoring, with two preset anomaly trigger conditions: First, potential anomalies, including the power frequency potential difference between any two accessible metal parts exceeding a set threshold, an instantaneous potential difference exceeding a preset instantaneous threshold with a duration meeting the trigger requirements, or a potential change rate exceeding a limit value; second, new noise events, including audio noise levels exceeding the allowable range compared to the reference value, or new noise peaks appearing in AC harmonics and radio frequency bands, with peak duration meeting the trigger requirements. Anomaly detection uses a combination of a sliding window averaging algorithm and a peak detection algorithm. The sliding window averaging calculation formula is as follows: The formula for window variance is: ,in The number of sampling points in the window. For the first Monitoring data from each sampling point For the first Window mean at time step Let Variance be the window variance, when When the data is deemed unstable, This is the window variance threshold. Peak detection uses an adaptive thresholding algorithm, and the threshold calculation formula is: , This is the sensitivity adjustment coefficient, when the monitoring data... Furthermore, if the duration exceeds the trigger hold time, it is determined to be an abnormal peak value. The above specific calculations achieve accurate anomaly detection, avoiding false triggers.

[0049] When the above-mentioned anomalies are detected, the interruption mechanism is immediately triggered, pausing the current normal monitoring process and automatically re-executing steps S2 to S4 to quickly respond to changes in operating conditions. During the re-execution process, step S2 will increase the data sampling density during the abnormal period to ensure accurate capture of abnormal features; step S3 will adjust the weight coefficients of the optimization function, prioritizing safety protection, increasing the optimization weight of the safety potential difference, and strengthening impedance optimization for newly added noise frequency bands; step S4 will accelerate the impedance adjustment rate to eliminate the anomaly as quickly as possible.

[0050] If the anomaly persists after repeatedly re-executing steps S2 to S4, the system automatically switches to emergency protection mode. In this mode, all variable impedance units switch to minimum impedance, maintaining a static equipotential network to ensure electrical safety. Simultaneously, an audible and visual alarm is activated, sending an alarm signal containing information such as the anomaly type and location. This signal can be uploaded to the host computer management system via Ethernet to notify maintenance personnel for timely handling. Continuous monitoring continues in emergency protection mode. Once the anomaly is resolved, the system automatically exits emergency mode and resumes the normal closed-loop optimization process. Furthermore, a manual trigger mechanism is provided, allowing maintenance personnel to issue commands via the host computer to force a re-execution of steps S2 to S4, facilitating proactive optimization of the system status after equipment maintenance or operational mode changes.

[0051] The aforementioned safety protection method for audio equipment first constructs and calibrates a static equipotential base network containing variable impedance units to obtain its reference impedance characteristics. Then, it synchronously and in real-time collects the potential data of each key node in the network and the final audio output signal data. Based on these data and the reference model, it constructs a multi-objective optimization function with the core objectives of minimizing dangerous power frequency potential differences (safe potential differences) and output noise. By solving this function, it dynamically generates and executes optimal control commands for each variable impedance unit, thereby actively and accurately adjusting the impedance distribution of the entire equipotential network in a wide frequency band. Ultimately, it achieves the dual technical effect of continuously suppressing dangerous potential differences between accessible components in the cabinet to ensure personal safety in complex operating environments, while effectively reducing audio noise introduced by grounding loops and interference to improve signal quality.

[0052] refer to Figure 2In one optional embodiment, based on the network reference impedance matrix, potential data, and audio data, a multi-objective optimization function is constructed with the goal of minimizing the safety potential difference and output audio noise. By solving the multi-objective optimization function, an optimal control instruction set for dynamically adjusting the impedance characteristics of each variable impedance unit is generated, including the following steps:

[0053] S11. Based on the network reference impedance matrix and the impedance regulation model of the variable impedance unit, establish a system node voltage prediction model with the control command set as the variable.

[0054] Specifically, the network reference impedance matrix, as an inherent characteristic of the static equipotential network, has its elements calibrated in the broadband step S1. This matrix is ​​the inverse of the nodal admittance matrix, denoted as... ,in To monitor the total number of nodes and key functional nodes, matrix elements Indicates the first The node and the first The inherent complex impedance between nodes, comprising resistive and reactive components, reflects the power transmission characteristics of the static network at different frequencies. The impedance regulation model of a variable impedance unit needs to be established based on its hardware structure; the actual impedance value of each variable impedance unit is determined by the control instruction set. Decision, among which The number of variable impedance units, For the first The control signal for a variable impedance unit (such as the voltage command corresponding to the PWM duty cycle) can be expressed as follows: In the formula In frequency Next The frequency response function of each variable impedance unit is determined by the parameters of the hardware itself, such as the MOSFET array and the high-frequency choke. For the first The mapping function between the control signal and the impedance value of each variable impedance unit was obtained through previous experimental calibration and satisfies... ,in and The first The minimum and maximum adjustable impedance values ​​of each variable impedance unit.

[0055] Based on the above, the system nodal voltage prediction model is established using the nodal voltage method. First, the total admittance matrix of the system, which includes variable impedance elements, is constructed. The process of constructing the total admittance matrix is ​​as follows: the network reference impedance matrix is ​​transformed into the reference admittance matrix. Then, based on the installation location of each variable impedance unit, its admittance value is determined. The values ​​are superimposed onto the corresponding elements of the reference admittance matrix to form the total admittance matrix. Assume the first... The variable impedance unit is connected to the first... The node and the first If there are nodes, then the elements in the total admittance matrix are... and Increase respectively ,element and Reduce respectively ,Right now:

[0056]

[0057] in, In frequency The total admittance matrix below, In frequency Next The admittance correction matrix corresponding to the variable impedance element is only applied to the _th variable impedance element. Line number Column, No. Line number Column, No. Line number Column, No. Line number The list contains non-zero elements, and all other elements are 0.

[0058] The final expression of the system node voltage prediction model is derived based on Kirchhoff's current law, that is, the algebraic sum of the node injected currents equals the product of the node admittance and the node voltage, in the following form:

[0059]

[0060] in, In frequency The predicted ground voltage vector for each monitoring node. For the first Predicted voltage to ground at each node (complex form, including amplitude and phase). The injected current vector for each monitoring node consists of the equipment operating current, leakage current, and electromagnetic induction coupling current, and can be obtained by collecting data from the current sensors at the monitoring nodes or by estimating based on the equipment's rated parameters and operating conditions. By solving this system of linear equations, a given set of control commands can be obtained. and frequency The predicted values ​​of the ground voltage at each node provide a basis for the subsequent quantification of safety and noise costs.

[0061] S12. Extract the instantaneous power frequency voltage values ​​of all monitoring nodes from the potential data, and calculate the set of measured power frequency potential differences between each pair of accessible metal parts based on the instantaneous power frequency voltage values; compare the set of measured power frequency potential differences with the preset safe voltage threshold to obtain the set of unsafe potential difference pairs that need to be optimized and eliminated.

[0062] Specifically, the extraction of instantaneous power frequency voltage values ​​requires the synchronous potential data collected in step S2. Considering the power frequency characteristics of the audio equipment power supply system, this section focuses on the voltage signal corresponding to the power frequency (determined according to the local power grid standards). Since the collected potential data may contain high-frequency interference, harmonic components, and random noise, targeted signal preprocessing is necessary. The preprocessing process includes: filtering the potential data using a second-order Butterworth bandpass filter, with the filter's center frequency set to the power frequency and a passband width of 5Hz (e.g., a 50Hz power frequency corresponds to a 47.5Hz-52.5Hz passband), to retain the fundamental power frequency signal and suppress other frequency components; subsequently, a moving average algorithm is used to smooth the filtered signal, with the sliding window size set to the number of sampling points within one cycle of the power frequency signal (e.g., a sampling rate of...). At that time, window size , (Using the power frequency), this eliminates glitches and interference in the signal, ensuring the accuracy of instantaneous voltage values.

[0063] The preprocessed potential data is the instantaneous sequence of power frequency voltage at each monitoring node, denoted as... ( , For monitoring node sequence number, The total number of monitoring nodes, (Sampling time), based on this sequence, the measured power frequency potential difference between each pair of accessible metal parts can be calculated. First, the corresponding monitoring nodes for all accessible metal parts are identified. Monitoring nodes for accessible parts such as the cabinet frame, the metal casings of each device, and the power distribution unit casing have all been set in step S2. Each accessible part uniquely corresponds to one monitoring node. Let the set of accessible parts be denoted as . ( The corresponding set of monitoring node numbers is For any two accessible parts and ( The corresponding instantaneous voltage values ​​of the monitoring nodes are as follows: and The instantaneous value of the power frequency potential difference between the two is The measured value of the power frequency potential difference is the effective value of that instantaneous value within one power frequency cycle. The calculation formula is as follows:

[0064]

[0065] in, t represents the power frequency period and t represents time.

[0066] For discrete sampled data, this effective value can be approximated by numerical integration, as shown in the formula:

[0067]

[0068] in, The sampling period is The number of sampling points within one power frequency cycle. For the first The instantaneous potential difference value corresponding to each sampling point.

[0069] The measured values ​​of the power frequency potential difference between all accessible metal parts constitute the set of measured values. Next, this set is compared with a preset safe voltage threshold. The setting is based on relevant electrical safety standards of the International Electrotechnical Commission (IEC) (such as IEC 61140) and audio equipment safety specifications. Considering the threshold characteristics of electric shock risk to the human body, it is usually set to AC 50V (RMS). This threshold is calculated based on the maximum safe current that the human body can pass through and the human body resistance (human body resistance can pass through...). It is usually estimated based on 1kΩ, corresponding to the safe current. ,Right now The comparison process employs an element-by-element comparison method, comparing each element in the set of measured values. ,like Then it is determined that the component is... There is a risk of electric shock, so it is included in the set of unsafe potential difference pairs. ;like If a component poses no risk of electric shock, it is not included in the set. This screening process identifies component pairs that require optimization to eliminate the risk of electric shock, providing specific optimization targets for the subsequent construction of safety cost sub-items.

[0070] S13. Perform a fast Fourier transform on the audio data to obtain the spectrum of the audio output signal. Extract all specific interference frequency components that exceed the preset background noise threshold and their corresponding amplitudes from the spectrum to form an interference frequency-amplitude feature set.

[0071] Specifically, the audio data (sampling rate) acquired based on step S2 The audio data (a 24-bit synchronous audio sequence) is first preprocessed to meet the requirements of FFT analysis. Since the audio data may contain the normal audio signal output by the device (the useful signal), and this step focuses on noise interference, the noise signal needs to be extracted first using signal separation technology—an adaptive noise cancellation algorithm is used. The input signal of the audio device is used as the reference signal, and the acquired output signal is used as the original signal. An adaptive filter adjusts the amplitude and phase of the reference signal to cancel out the useful components in the original signal, outputting only a signal containing noise interference. ; then the noise signal Window function processing is performed, using the Hanning window to weight the signal. The window function expression can be: ( ),in The number of sampling points for the FFT is determined by the principle that it satisfies the Nyquist sampling theorem and is an integer power of 2 (e.g., ...). (or 16384), the function of the window function is to reduce leakage of the signal spectrum and ensure the accuracy of spectrum analysis.

[0072] Performing an FFT operation on the windowed noise signal transforms the time-domain noise signal into a frequency-domain spectrum. The mathematical expression for FFT is:

[0073]

[0074] in, For the first The complex spectral values ​​at each frequency point contain the amplitude and phase information of that frequency point. The sampling period for audio data. This is the frequency point index, and the corresponding actual frequency is... The spectrum coverage obtained through FFT operation is from 0 to... (Nyquist frequency) covers the effective frequency range of audio signals (20Hz-20kHz) and possible radio frequency interference frequencies (such as radio frequency bands below 1MHz).

[0075] The noise amplitude at each frequency point is calculated based on the frequency domain spectrum. The amplitude calculation uses the modulus operation of complex numbers, and the formula is as follows:

[0076]

[0077] in, For the first Noise amplitude at each frequency point and These are the complex spectrum values. The real and imaginary parts. To accurately identify interference frequency components, a preset noise floor threshold needs to be established. The threshold is set as follows: when the audio device is running unloaded (without input signal), its output noise signal is collected and FFT analysis is performed to calculate the noise amplitude at each frequency point. The maximum value of the amplitude at each frequency point is taken as the background noise threshold, or the mean of the amplitude plus 3 times the standard deviation is taken as the threshold using a statistical method to ensure that the background noise and external interference noise can be effectively distinguished.

[0078] The process of constructing the interference frequency-amplitude feature set is as follows: traverse all frequency points after FFT, and for each frequency point... If its corresponding noise amplitude If the frequency is determined to be an interference frequency, then the actual frequency value of that frequency point is determined. and the corresponding noise amplitude Compositional features All such features constitute the set of interference frequency-amplitude features. The interference frequency components in this set can include: power frequency harmonics (such as the second and third harmonics of 50Hz, 100Hz, 150Hz, etc.), electromagnetic interference frequencies coupled from power transformers, and radio frequency interference frequencies (such as walkie-talkie communication frequencies, radiation frequencies from industrial equipment, etc.). These frequency components are coupled to the audio signal path through ground loops and are the main source of audio noise. Their amplitude directly reflects the severity of noise interference, providing key parameters for the subsequent quantification of noise cost sub-items.

[0079] S14. Under a given set of control instructions, based on the system node voltage prediction model, predict the power frequency potential difference between each pair of components in the unsafe potential difference pair set, and predict the ground voltage of the audio signal star ground bus at each frequency point in the interference frequency-amplitude characteristic set.

[0080] Specifically, the calculation of the predicted power frequency potential difference focuses on the set of unsafe potential difference pairs obtained in S12. Each unsafe component is numbered according to its corresponding monitoring node. ( For components monitoring nodes For components (Monitoring nodes). Based on the node voltage prediction model established in S11, under a given set of control commands... Under these conditions, the power frequency can be obtained by solving. Predicted ground voltage values ​​of each node and (In complex form), the predicted power frequency potential difference between the two is the difference in the complex magnitude of the two predicted voltage values. The calculation formula is as follows:

[0081]

[0082] The physical meaning of this formula is that the potential difference between accessible metal parts is determined by the voltage difference between their corresponding monitoring nodes and ground. Since all accessible parts are connected through an equipotential network, their potential difference is essentially the manifestation of node voltage under the impedance distribution of the grounding network. A predictive model can accurately quantify the changing trend of this potential difference under different control commands, thereby determining the impact of control commands on the risk of electric shock. For unsafe potential differences affecting the set... All component pairs are calculated using this formula to determine their corresponding predicted power frequency potential differences, forming a set of predicted values. .

[0083] The calculation of the predicted ground voltage is based on the interference frequency-amplitude characteristic set obtained in S13. The focus is on the voltage to ground of the audio signal star ground bus, as this bus is the common connection point for the signal grounds of all audio devices. Changes in its voltage to ground will be introduced into the audio signal path through the signal ground leads, directly affecting the noise level of the audio output. First, identify the monitoring node corresponding to the audio signal star ground bus (already set as the midpoint or two end nodes of the bus in step S1, denoted as node). For feature sets Each interference frequency point Substituting this into the node voltage prediction model of S11, the node voltage at this frequency is obtained. Predicted value of voltage to ground (In complex form), its amplitude for This amplitude reflects the star-grounded bus at the interference frequency. The greater the potential fluctuation, the stronger the noise that will interfere with the audio signal through the signal ground. It should be noted that the interference frequency point... This may cover the mid-to-high frequency range (e.g., 1kHz-1MHz), at which point the total admittance matrix in the node voltage prediction model... The impedance characteristics of the variable impedance element in the mid-to-high frequency range need to be considered (e.g., the inductive reactance of a high-frequency choke increases with frequency) to ensure the accuracy of the predicted values. For the feature set... For all interference frequency points, the corresponding predicted value of the voltage amplitude to ground is calculated according to this formula, forming a set of predicted values. .

[0084] During the prediction process, it is necessary to ensure the control instruction set Consistency is crucial. For the same set of control commands, both power frequency potential difference prediction and ground voltage prediction at all interference frequency points must be used simultaneously to ensure that the prediction results reflect the comprehensive impact of the control command set on safety and noise. Furthermore, the prediction model should employ efficient numerical calculation methods (such as LU decomposition) to avoid computational inefficiencies caused by high matrix dimensions (e.g., when the number of monitoring nodes is large), ensuring rapid convergence of the optimization process. Through this step of prediction, the mapping relationship between the control command set and safety risks and noise interference is quantified, providing clear input variables and output indicators for the subsequent construction of multi-objective optimization functions.

[0085] S15. Construct a multi-objective optimization function based on the predicted power frequency potential difference and the predicted voltage to ground. The multi-objective optimization function includes a safety cost sub-item and a noise cost sub-item. The safety cost sub-item is calculated based on the deviation between the predicted power frequency potential difference and the safety voltage threshold to quantify the risk of electric shock. The noise cost sub-item is calculated based on the predicted voltage to ground and the transfer function of the corresponding audio output terminal to quantify the noise interference level.

[0086] Specifically, the safety cost sub-item is constructed based on the set of predicted power frequency potential differences obtained in S14. Its core principle is to reflect the magnitude of electric shock risk by quantifying the deviation between the predicted chemical frequency potential difference and the safe voltage threshold—the greater the deviation, the higher the electric shock risk and the higher the safety cost. Considering the combined impact of multiple unsafe potential difference pairs, the safety cost sub-item is constructed using a weighted sum of squares, as shown in the formula:

[0087]

[0088] in, For the safety cost sub-item, For unsafe potential difference pairs The weighting coefficients are determined based on factors such as the accessibility probability, contact area, and location of the component. Components with high accessibility probability (e.g., front door of the cabinet, equipment panel) and large contact area have larger weighting coefficients (e.g., 0.8-1.0), while components with low accessibility probability (e.g., back of the cabinet, internal equipment connection points) have smaller weighting coefficients (e.g., 0.2-0.5). All weighting coefficients satisfy... To ensure that the magnitude of safety cost sub-items is consistent; This means that the cost is calculated only when the predicted potential difference exceeds the safety threshold; when it does not exceed the threshold, the cost contribution is 0, which is consistent with the threshold characteristics of electric shock risk. The introduction of the squared term aims to amplify the cost contribution of larger deviations, prompting the optimization algorithm to prioritize the elimination of severely out-of-range potential difference pairs.

[0089] The noise cost sub-item is constructed based on the set of predicted ground voltage values ​​obtained in S14. The core of this approach is to quantify the impact of the predicted ground voltage of the star-grounded bus on the audio output after it is transmitted through the signal path, reflecting the noise interference level—the larger the transmitted noise amplitude, the more severe the noise interference, and the higher the noise cost. First, it is necessary to determine the transfer function of the star-grounded bus's ground voltage to the audio output terminal. This transfer function characterizes the ground voltage noise at the interference frequency. The gain of the audio output is obtained by injecting a test voltage signal with known amplitude and frequency into the star-grounded busbar, while simultaneously acquiring the corresponding noise amplitude at the audio output. The amplitude of the transfer function is the ratio of the output noise amplitude to the injected test voltage amplitude, and the phase is the phase difference between the two. (in plural form), where For the injected test voltage, This represents the noise voltage at the audio output. Transfer function. It is necessary to consider the interference frequency-amplitude characteristic set. The calibration is performed at all frequency points and stored as a frequency-gain mapping table for querying and calling during optimization calculations.

[0090] Based on the transfer function, the noise cost sub-item is constructed using a weighted sum, as shown in the formula:

[0091]

[0092] in, For noise cost sub-item, Interference frequency point The weighting coefficients are determined based on the degree of influence of the frequency component on the audio quality: interference frequencies in frequency bands sensitive to human hearing (such as 1kHz-4kHz) have larger weighting coefficients (such as 0.7-1.0), while high-frequency or low-frequency interference frequencies that are not sensitive to human hearing have smaller weighting coefficients (such as 0.1-0.4). All weighting coefficients satisfy the following conditions: ; For the transfer function at frequency The amplitude below; It represents the predicted noise amplitude at the audio output after the ground voltage noise is transmitted. The introduction of the square term is also to amplify the cost contribution of larger noise amplitudes, prompting the optimization algorithm to prioritize the suppression of strong interference frequency components.

[0093] The final form of the multi-objective optimization function is a weighted sum of the safety cost sub-item and the noise cost sub-item, as shown in the formula:

[0094]

[0095] in, The combined value of the multi-objective optimization function. The weighting factor for balancing safety costs and noise costs ( Its value can be dynamically adjusted according to real-time operating conditions: when the grid voltage fluctuates greatly or the equipment leakage current detection value is too high, increase the value. Prioritize safety; when the audio signal is weak (such as a microphone input signal) and extremely high sound quality is required, reduce... Prioritize noise suppression; under normal operating conditions, it can be set... This achieves a balanced optimization of safety and sound quality. The physical meaning of this optimization function is that by comprehensively calculating the quantitative costs of electric shock risk and noise interference, the multi-objective optimization problem is transformed into a single-objective weighted sum optimization problem, which is convenient to solve using numerical optimization algorithms. At the same time, through the dynamic adjustment of the weight coefficients, the optimization objective can be adaptively adjusted under different operating conditions.

[0096] S16. With the goal of minimizing the value of the multi-objective optimization function, a numerical optimization algorithm is used to search for the optimal control instruction set within the adjustable impedance range of the variable impedance unit; whereby the optimal control instruction set represents the combination of control instructions that minimizes the predicted risk of electric shock and the level of noise interference.

[0097] Specifically, the adjustable impedance range of a variable impedance unit needs to be determined based on its hardware characteristics, and each variable impedance unit has its minimum adjustable impedance. and maximum adjustable impedance (Determined by the on-resistance of the MOSFET array, the range of the precision resistor network, and the parameters of the high-frequency choke, etc.), the corresponding control command set. There is also a corresponding adjustable range. ( , This represents the minimum commanded control value corresponding to the i-th variable impedance unit. The range corresponding to the maximum command adjustment value of the i-th variable impedance unit can be obtained through preliminary experimental calibration (e.g., by changing the control command and measuring the actual impedance value, establishing a mapping relationship between the control command and the impedance value, and thus determining the adjustable boundary of the command). In addition, the optimization process also needs to consider other constraints, such as the power capacity constraint of the variable impedance unit (to avoid damage due to excessive power loss caused by insufficient impedance under high current conditions), and the response speed constraint (the impedance adjustment corresponding to the control command must be completed within the system's allowed time to avoid affecting the real-time performance of the audio signal). These constraints can be incorporated into the optimization process through inequalities, denoted as... ( , (Number of constraints).

[0098] The selection of a numerical optimization algorithm must balance optimization accuracy and computational efficiency. Considering that the control instruction set is a continuous variable and the optimization function may have local optima, the Particle Swarm Optimization (PSO) algorithm is chosen here. This algorithm is robust, converges quickly, and is easy to implement, making it suitable for multi-constraint optimization problems in continuous space. The core idea of ​​the PSO algorithm is to simulate the foraging behavior of a flock of birds, and to search for the optimal solution by updating the position and velocity of particles. The specific implementation steps are as follows:

[0099] 1) Initialize the particle swarm: Set the particle swarm size to... (Typically a value of 50-100), the position of each particle corresponds to a set of control instructions. ( The initialization of particle positions uses random sampling, generating them randomly within the adjustable range of control commands to ensure that the initial particle swarm can uniformly cover the entire search space; the velocity of each particle... Also initialized randomly (n is the number of particles), the velocity range is set to... ( (Maximum permissible speed).

[0100] 2) Calculate the fitness value: for each particle's corresponding control instruction set Substitute into the multi-objective optimization function constructed in S15 The fitness value of the particle (i.e., the comprehensive value of the optimization function) is calculated. The smaller the fitness value, the better the overall performance of the control instruction set. Simultaneously, it is checked whether the particle satisfies all constraints. If the conditions are not met, the fitness value is penalized (e.g., multiplied by a large penalty coefficient) to ensure that the optimization algorithm prioritizes searching for solutions that satisfy the constraints.

[0101] 3) Update individual optimal and global optimal: Each particle updates its current fitness value with its own historical best fitness value (individual optimal). The optimal position of an individual particle is updated to its current position by comparing its fitness value with the fitness value of all particles. The optimal position corresponding to the best fitness value is then selected as the global optimal position. .

[0102] 4) Update particle position and velocity: Update each particle's position and velocity according to the velocity update formula and position update formula of the PSO algorithm. The velocity update formula is:

[0103]

[0104] The position update formula is:

[0105]

[0106] in, This represents the number of iterations. The inertial weight is used to balance the global search and local search capabilities. A larger value is taken in the early stage of the iteration to enhance the global search, and a smaller value is taken in the later stage of the iteration to enhance the local search. and The acceleration coefficient represents the particle's ability to learn from its own historical best and global best, respectively. and Use a random number between 0 and 1 to increase the randomness of the search.

[0107] 5) Iteration termination judgment: Set the maximum number of iterations. (Typically a value of 100-200) or the change in minimum fitness value (like When the number of iterations reaches Or the change in fitness value over multiple consecutive iterations is less than When the algorithm converges, the iteration stops.

[0108] The global optimal position after iterative convergence The corresponding control instruction set is the optimal control instruction set. This instruction set enables multi-objective optimization functions The minimum value represents the minimum combined predicted risk of electric shock and noise interference. To ensure the reliability of the optimal solution, the optimal control instruction set needs to be verified: It is substituted into the system node voltage prediction model, and the safety cost and noise cost sub-items are recalculated to confirm that the combined value meets the design requirements. Simultaneously, hardware-in-the-loop simulation or actual equipment testing is used to verify whether the impedance adjustment effect corresponding to this instruction set meets expectations. If deviations exist, the parameters of the optimization algorithm (such as particle swarm size, number of iterations, weight coefficients, etc.) can be adjusted appropriately to re-search until the optimal control instruction set that meets the requirements is obtained. This optimal control instruction set will serve as the input to step S4, driving the variable impedance unit to perform impedance adjustment, achieving a dynamic balance between safety protection and sound quality optimization.

[0109] In an optional embodiment, the expression for the security cost sub-item is:

[0110]

[0111] in, This represents the safety cost sub-item. This represents the control command vector consisting of the control commands for all variable impedance units. This represents the set of unsafe potential difference pairs. This represents an unsafe potential difference between a pair of accessible metal parts in the set. Indicates in the control command vector Lower component pair Predicted value of the power frequency potential difference between them; This indicates the safe voltage threshold.

[0112] Specifically, the core logic of the expression for the safety cost sub-item is as follows: iterate through all accessible metal parts that pose a risk of electric shock, quantify the cost only for the portion of the predicted power frequency potential difference that exceeds the safety threshold, and amplify the cost contribution of the deviation exceeding the standard through squaring operations, prompting the optimization algorithm to prioritize the elimination of severely excessive potential differences, while ignoring the risk-free contribution of the portion that does not exceed the standard, thereby achieving accurate quantification and optimization guidance of electric shock risk.

[0113] The safety cost sub-item is a core indicator for quantifying the electric shock risk of a system. Its value is positively correlated with the electric shock risk; that is, the higher the value, the greater the risk of electric shock in the system, requiring adjustments to the control command vector to reduce this value until safety requirements are met. As an important component of the multi-objective optimization function, this indicator, together with the noise cost sub-item, constitutes the optimization objective, guiding the search for the optimal control command. This represents the control command vector, composed of control commands from all variable impedance units. Each variable impedance unit corresponds to one control command (such as PWM duty cycle, voltage command, etc.), and the vector dimension is the same as the number of variable impedance units. ( (where the number of variable impedance units is 1), For the first The control commands for each variable impedance unit. This vector is the core variable of the optimization algorithm; by adjusting its value, the impedance value of each variable impedance unit can be changed, thereby affecting the system node voltage and the potential difference between components. This represents the set of unsafe potential difference pairs, specifically the set of accessible metal component pairs selected in the preceding steps whose measured power frequency potential difference exceeds the safe voltage threshold. This set clarifies the quantification range of the safety cost sub-item, calculating only component pairs posing a risk of electric shock, avoiding invalid calculations for risk-free component pairs, and improving optimization efficiency. Elements in the set... Each component must uniquely correspond to an accessible metal part in the system, and the potential difference only needs to be calculated once per component (no need for repeated calculations). ). This represents a pair of accessible metal parts in the set Ω representing unsafe potential differences, where , These represent two distinct accessible metal components (such as the rack frame and equipment housing, or the equipment housing and power distribution unit housing). Note that the components... and Essentially, they are the same pair of components with equal absolute potential differences, so the expression only needs to iterate once to avoid redundant calculations. Indicates in the control command vector Under the action, the component The predicted value of the power frequency potential difference between the nodes. This value is obtained by solving the system node voltage prediction model, that is, based on the control command vector. Determine the impedance value of each variable impedance element, construct the total admittance matrix of the system, and solve for the component using Kirchhoff's current law. , The absolute value of the difference between the voltage to ground at the corresponding monitoring node and the voltage at the monitoring node is the predicted value, and its unit is consistent with the safe voltage threshold (both are effective values ​​in volts). This value directly reflects the level of electric shock risk to the component pair under a specific control command and is the core input parameter of the safety cost sub-item. The safe voltage threshold is a critical value that defines the risk of electric shock. It is set according to relevant electrical safety standards of the International Electrotechnical Commission (IEC) and safety specifications for audio equipment, and is typically taken as 50V AC (RMS). This threshold is based on the maximum safe current that the human body is allowed to pass through. (Usually estimated at 50mA) and human body resistance (Usually estimated at 1kΩ) Calculated to obtain ( This ensures that when the body is below this threshold, the parts in contact with the human body will not cause electric shock.

[0114] In an optional embodiment, the expression for the noise cost sub-item is:

[0115]

[0116] in, This represents the noise cost sub-item. This represents a control command vector consisting of control commands for all variable impedance units. This represents a set of reference interference frequencies extracted from the interference frequency-amplitude feature set; Represents the first in the reference interference frequency set One reference interference frequency; For frequency The corresponding angular frequency; Indicates in the control command vector Below, the audio signal star-grounded bus is at frequency Predicted voltage to ground at the location; This indicates the frequency at which the audio signal flows from the star-grounded bus to the audio output of the power amplifier. The preset transfer function value at the location; This represents the set of all specific interference frequencies extracted from the interference frequency-amplitude feature set; Represents the first in a specific set of interference frequencies A specific interference frequency; Represents a control command vector Down-prediction, audio output at frequency A reference quantity that is positively correlated with the noise amplitude at that location; These are the preset weighting coefficients.

[0117] Specifically, the expression for the noise cost sub-item adopts a two-term structure of "baseline interference quantization + specific interference quantization". The core logic is: for the core base interference frequency that affects sound quality, the noise contribution is quantified by coupling the ground voltage and the transfer function; for the specific interference frequency that needs to be focused on, the quantization is supplemented by the reference quantity of the associated noise amplitude; and then the weighting coefficient is used to adjust the proportion of the two contributions to achieve comprehensive and accurate quantification of noise interference, providing a clear noise suppression guide for the optimization algorithm.

[0118] Noise cost sub-item It is a core indicator for quantifying the noise interference level at the audio output. Its value is positively correlated with the noise interference intensity; the higher the value, the more severe the audio output noise, requiring adjustment of the control command vector. Lowering this value balances safety performance and audio quality. This indicator, along with the safety cost sub-item, constitutes a multi-objective optimization function, jointly guiding the search for the optimal control command vector. Control command vector Consistent with the definition of the control command vector in the safety cost sub-item, it consists of the control commands for all variable impedance units, i.e. ( (The number of variable impedance units). By adjusting... The value of can change the impedance characteristics of each variable impedance unit, thereby affecting the voltage to ground of the audio signal star-grounded bus and the noise amplitude at the audio output. It is a core variable for controlling noise costs. Reference interference frequency set This set, selected from the previously extracted set of interference frequency-amplitude features, includes core interference frequencies (such as power frequency harmonics and characteristic interference frequencies of power coupling) that have the most significant impact on audio quality and occur most frequently. This set focuses on key noise sources, avoiding computational redundancy caused by full frequency traversal and improving optimization efficiency. Represents the set of reference interference frequencies The first in The reference interference frequency, whose value directly corresponds to the frequency point with higher amplitude and greater impact on sound quality in the interference frequency-amplitude feature set, is the core object of noise quantization. Indicates the reference interference frequency The corresponding angular frequency is used to match the transfer function. The frequency parameter format is as follows. The transfer function takes angular frequency as an input variable. This conversion enables precise matching between the interference frequency and the transfer function, ensuring operational consistency. Indicates in the control command vector Under its influence, the audio signal star-grounded bus at the reference interference frequency... The predicted voltage to ground at the location is obtained by solving the system node voltage prediction model. It reflects the potential fluctuation of the ground bus at a specific interference frequency. The greater the fluctuation, the stronger the noise entering the audio output terminal. It is the core input parameter for noise quantization. This indicates the frequency of the audio signal from the star-grounded bus to the audio output of the power amplifier at an angular frequency. The preset transfer function value at the location. This transfer function is a fixed value after prior calibration, which represents the gain coefficient of the ground voltage noise of the ground bus at the corresponding angular frequency, which is fed into the audio output terminal through the signal path. Its amplitude directly determines the degree of influence of ground voltage fluctuation on the output noise, and the phase information is incorporated into the overall quantization through complex modulus calculation. This represents a specific set of interference frequencies, also extracted from the interference frequency-amplitude feature set, including special interference frequencies that require key monitoring (such as non-common interference frequencies like radio frequency communication interference and internal coupling interference within equipment). This set supplements key noise sources beyond the reference interference frequencies, achieving full coverage quantification of noise interference. Represents a specific set of interference frequencies The first in A specific interference frequency, the value of which is determined for a specific interference source in a given scenario, can be dynamically adjusted according to the equipment's working environment and conditions to ensure effective quantification of sudden or special noise sources. Representation and control command vector Under the influence of this, the audio output terminal at a specific interference frequency A reference value that is positively correlated with the noise amplitude. This reference value can be obtained by multiplying the measured noise amplitude value by the calibration coefficient, or by directly extracting it from the audio output signal spectrum. Its value variation trend is completely consistent with the actual noise amplitude, and it is used to quantify the impact of a specific interference frequency on the audio output. This represents the preset weighting coefficient, with a value range of [value range missing]. This is used to adjust the contribution ratio of the reference interference frequency quantization term to the specific interference frequency quantization term. When the reference interference (such as power frequency harmonics) is the main noise source, it can be reduced. When specific interference (such as radio frequency interference) has a significant impact, it can be increased. This enables the prioritization of noise suppression to be adapted under different operating conditions.

[0119] In an optional embodiment, the multi-objective optimization function further includes a power consumption cost sub-term, which is used to constrain the variation range of the control command; the expression for the power consumption cost sub-term is:

[0120]

[0121] in, This indicates the cost of regulating power consumption. This represents a control command vector consisting of control commands for all variable impedance units. Represents the control command vector The Middle New control commands for a variable impedance unit; This indicates the prediction based on the impedance control model, in the control command. Next The complex impedance value of each variable impedance element at the target operating frequency; Indicates the first The current complex impedance value of each variable impedance unit before adjustment; This represents the total number of variable impedance units.

[0122] Specifically, this embodiment adds a sub-item for regulating power consumption cost to the multi-objective optimization function. By quantifying the impedance change amplitude of the variable impedance unit to constrain the adjustment range of the control command, a balance between "safety, noise, and power consumption" is achieved. This sub-item overcomes the limitations of focusing only on safety and noise, avoiding impedance abrupt changes caused by large fluctuations in control commands, thereby reducing device regulation power consumption and component losses. At the same time, it ensures the stability of the optimization command and the reliability of device operation. Together with the aforementioned safety cost and noise cost sub-items, it forms a more comprehensive multi-objective optimization system.

[0123] The core logic of the expression for regulating the power consumption cost sub-item is: iterate through all variable impedance units, quantify the deviation between the predicted impedance after adjustment and the current actual impedance of each unit, and convert the impedance change magnitude into power consumption cost through the sum of squares operation. The larger the deviation, the higher the cost. This forces the optimization algorithm to minimize the impedance change as much as possible when adjusting the control command, thereby constraining the regulation of power consumption and equipment loss, and achieving a multi-dimensional balance of the optimization goal.

[0124] The power consumption cost sub-item is a core indicator for quantifying the power consumption and variation range of the variable impedance unit. Its value is positively correlated with the impedance variation range; a larger value indicates more drastic impedance adjustment, resulting in higher power consumption and a greater risk of component loss. This sub-item serves as a supplement to the multi-objective optimization function, working synergistically with the safety cost and noise cost sub-items to ensure that the optimization result satisfies both safety and sound quality requirements while also considering low power consumption and device stability. This represents the control command vector, which consists of the control commands for all variable impedance units. This vector is the core carrier of the correlation between impedance change and power consumption cost. Its value changes directly determine the impedance adjustment range of each variable impedance unit, thereby affecting the control power consumption cost. Represents the control command vector The Middle A new control instruction for each variable impedance unit. This instruction is the command to be executed obtained by the optimization algorithm. Through the preceding impedance control model, it can be transformed into the corresponding predicted impedance value, which is the core input for calculating the impedance change amplitude. This indicates the prediction based on the impedance control model under the new control command. Under the action, the first The complex impedance value of each variable impedance unit at the target operating frequency. This impedance value includes resistive and reactive components, and is consistent with the impedance parameter format in the node voltage prediction model. Its prediction is based on the previously established impedance control model (i.e., the mapping relationship between control commands and impedance). The target operating frequency covers the power frequency and the core interference frequency to ensure that the predicted value fits the actual working scenario. Indicates the first The current complex impedance value of each variable impedance unit before adjustment. This value is the measured impedance during real-time operation of the device, acquired by an impedance sensor, and serves as a benchmark for measuring the magnitude of impedance change. Its numerical stability directly affects the quantitative accuracy of controlling power consumption and cost. This represents the total number of variable impedance units, and the control command vector. The dimensions are consistent. Traversing this total number can cover all variable impedance units, ensuring that the impedance change of each unit is included in the power consumption cost quantification, and avoiding a sharp increase in overall power consumption due to large adjustments of individual units.

[0125] The aforementioned safety protection method for audio equipment first constructs and calibrates a static equipotential base network containing variable impedance units to obtain its reference impedance characteristics. Then, it synchronously and in real-time collects the potential data of each key node in the network and the final audio output signal data. Based on these data and the reference model, it constructs a multi-objective optimization function with the core objectives of minimizing dangerous power frequency potential differences (safe potential differences) and output noise. By solving this function, it dynamically generates and executes optimal control commands for each variable impedance unit, thereby actively and accurately adjusting the impedance distribution of the entire equipotential network in a wide frequency band. Ultimately, it achieves the dual technical effect of continuously suppressing dangerous potential differences between accessible components in the cabinet to ensure personal safety in complex operating environments, while effectively reducing audio noise introduced by grounding loops and interference to improve signal quality.

[0126] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0127] Based on the same inventive concept, this application also provides a system for implementing the aforementioned audio device security protection method. The solution provided by this system is similar to the implementation described in the above method; therefore, the specific limitations in one or more audio device security protection system embodiments provided below can be found in the limitations of the audio device security protection method described above, and will not be repeated here.

[0128] In one exemplary embodiment, such as Figure 3 As shown, a security protection system 30 for audio devices is provided to implement the methods in the above-described method embodiments. The system includes:

[0129] The network reference construction module 31 is used to establish a static equipotential base network for audio equipment, perform broadband impedance parameter calibration on the static equipotential base network, and obtain the network reference impedance matrix; wherein, the main grounding reference point, the audio signal star grounding bus, and the preset positions of the potential connection architecture in the static equipotential base network are equipped with variable impedance units.

[0130] The real-time monitoring and acquisition module 32 is used to synchronously sample the ground potential of multiple monitoring nodes set on the static equipotential foundation network to obtain potential data; and to synchronously acquire the audio output signal of the power amplifier of the audio equipment to obtain audio data.

[0131] The multi-objective optimization decision module 33 is used to construct a multi-objective optimization function based on the network reference impedance matrix, potential data, and audio data, with the optimization objectives of minimizing the safety potential difference and output audio noise; by solving the multi-objective optimization function, an optimal control instruction set is generated for dynamically adjusting the impedance characteristics of each variable impedance unit; wherein, the safety potential difference represents the power frequency potential difference between any two accessible metal parts in the audio device that can cause the risk of electric shock.

[0132] The impedance dynamic adjustment module 34 is used to drive each variable impedance unit to adjust its impedance value over a wide frequency band according to the optimal control instruction set.

[0133] The adaptive feedback monitoring module 35 is used to continuously monitor the ground potential and noise of the audio output signal of the monitoring node. When an abnormal potential or a new noise event is detected, the real-time monitoring and acquisition module, the multi-objective optimization decision-making module and the impedance dynamic adjustment module are activated to re-execute the corresponding operations.

[0134] Embodiments of this application also provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the aforementioned method embodiments.

[0135] Embodiments of this application also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method embodiments.

[0136] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0137] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A security protection method of an audio device, characterized by, The method includes: S1. Establish a static equipotential base network for the audio equipment, perform broadband impedance parameter calibration on the static equipotential base network, and obtain the network reference impedance matrix; wherein, the main grounding reference point, the audio signal star grounding bus, and the preset positions of the potential connection architecture in the static equipotential base network are provided with variable impedance units. S2. The ground potential of multiple monitoring nodes set on the static equipotential foundation network is synchronously sampled to obtain potential data; the audio output signal of the power amplifier of the audio device is synchronously acquired to obtain audio data. S3. Based on the network reference impedance matrix, the potential data, and the audio data, construct a multi-objective optimization function with the optimization objective of minimizing the safety potential difference and output audio noise; by solving the multi-objective optimization function, generate an optimal control instruction set for dynamically adjusting the impedance characteristics of each variable impedance unit; wherein, the safety potential difference represents the power frequency potential difference between any two accessible metal parts in the audio device that could cause the risk of electric shock; S4. Drive each of the variable impedance units to adjust its impedance value in a wide frequency band according to the optimal control instruction set; S5. Continuously monitor the ground potential of the monitoring node and the noise of the audio output signal. When an abnormal potential or a new noise event is detected, repeat S2 to S4.

2. The method of claim 1, wherein, Based on the network reference impedance matrix, the potential data, and the audio data, a multi-objective optimization function is constructed with the goal of minimizing the safety potential difference and the output audio noise. By solving the multi-objective optimization function, an optimal set of control instructions is generated for dynamically adjusting the impedance characteristics of each variable impedance unit, including: S11. Based on the network reference impedance matrix and the impedance regulation model of the variable impedance unit, establish a system node voltage prediction model with the control command set as the variable. S12. Extract the instantaneous power frequency voltage values ​​of all monitoring nodes from the potential data, and calculate the set of measured power frequency potential differences between each pair of accessible metal parts based on the instantaneous power frequency voltage values; compare the set of measured power frequency potential differences with a preset safe voltage threshold to obtain a set of unsafe potential difference pairs that need to be optimized and eliminated. S13. Perform a fast Fourier transform on the audio data to obtain the spectrum of the audio output signal, and extract all specific interference frequency components that exceed the preset background noise threshold and their corresponding amplitudes from the spectrum to form an interference frequency-amplitude feature set. S14. Under a given set of control instructions, based on the system node voltage prediction model, predict the power frequency potential difference between each pair of components in the unsafe potential difference pair set, and predict the ground voltage of the audio signal star ground bus at each frequency point in the interference frequency-amplitude characteristic set. S15. Construct the multi-objective optimization function based on the predicted power frequency potential difference and the predicted voltage to ground; wherein, the multi-objective optimization function includes a safety cost sub-item and a noise cost sub-item, the safety cost sub-item is calculated based on the deviation between the predicted power frequency potential difference and the safety voltage threshold, and is used to quantify the risk of electric shock; the noise cost sub-item is calculated based on the predicted voltage to ground and the transfer function of the corresponding audio output terminal, and is used to quantify the noise interference level; S16. With the goal of minimizing the value of the multi-objective optimization function, a numerical optimization algorithm is used to search for the optimal control instruction set within the adjustable impedance range of the variable impedance unit; wherein, the optimal control instruction set represents the combination of control instructions that minimizes the predicted risk of electric shock and the level of noise interference.

3. The method of claim 2, wherein, The expression for the security cost sub-item is: wherein, represents the safe cost sub-item, represents a control instruction vector composed of control instructions of all the variable impedance units, represents the unsafe potential difference pair set, represents a pair of accessible metal components in the unsafe potential difference pair set, represents the power frequency potential difference prediction value between the control instruction vector lower component pair ; represents the safe voltage threshold.

4. The method of claim 2, wherein, The expression for the noise cost sub-item is: wherein, represents the noise cost subterm, represents a control instruction vector composed of control instructions of all the variable impedance units; represents a set of reference interference frequency set extracted from the interference frequency-amplitude characteristic set; represents the th reference interference frequency in the reference interference frequency set; is the frequency corresponding angular frequency; represents the ground voltage prediction value of the audio signal star ground bus at frequency under control instruction vector ; represents the preset transfer function value of the audio output end of the power amplifier from the audio signal star ground bus at angular frequency ; represents all specific interference frequency sets extracted from the interference frequency-amplitude characteristic set; represents the th specific interference frequency in the specific interference frequency set; represents a reference quantity positively correlated with the noise amplitude of the audio output end at frequency predicted under control instruction vector ; is a preset weight coefficient.​​ 5. The method according to any one of claims 2 to 4, characterized in that, The multi-objective optimization function also includes a power consumption cost sub-term, which is used to constrain the variation range of the control command; the expression for the power consumption cost sub-term is: wherein, represents the control instruction vector composed of control instructions of all the variable impedance units; represents the control instruction vector composed of control instructions of all the variable impedance units; represents the new control instruction of the th variable impedance unit in the control instruction vector represents the complex impedance value of the th variable impedance unit at the target operating frequency under the control instruction represents the current complex impedance value of the th variable impedance unit before adjustment; is the total number of the variable impedance units.​​​​​ 6. A security protection system for an audio device, used to implement the method according to any one of claims 1 to 5, characterized in that, The system includes: The network reference construction module is used to establish a static equipotential base network for audio devices, perform broadband impedance parameter calibration on the static equipotential base network, and obtain the network reference impedance matrix; wherein, the main grounding reference point, the audio signal star grounding bus, and the preset positions of the potential connection architecture in the static equipotential base network are provided with variable impedance units. The real-time monitoring and acquisition module is used to synchronously sample the ground potential of multiple monitoring nodes set on the static equipotential foundation network to obtain potential data; and to synchronously acquire the audio output signal of the power amplifier of the audio device to obtain audio data. A multi-objective optimization decision module is used to construct a multi-objective optimization function based on the network reference impedance matrix, the potential data, and the audio data, with the optimization objectives being to minimize the safety potential difference and the output audio noise; by solving the multi-objective optimization function, an optimal control instruction set is generated for dynamically adjusting the impedance characteristics of each of the variable impedance units; wherein, the safety potential difference represents the power frequency potential difference between any two accessible metal parts within the audio device that could cause an electric shock risk; An impedance dynamic adjustment module is used to drive each of the variable impedance units to adjust its impedance value over a wide frequency band according to the optimal control instruction set. The adaptive feedback monitoring module is used to continuously monitor the ground potential of the monitoring node and the noise of the audio output signal. When an abnormal potential or a new noise event is detected, the real-time monitoring and acquisition module, the multi-objective optimization decision-making module, and the impedance dynamic adjustment module are activated to re-execute the corresponding operations.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 5.

Images