An adaptive miniature circuit breaker switch status indication system

Abstract

The application discloses a self-adaptive miniature circuit breaker switch state indication system, comprising: a mechanical structure unit; a sensor network unit; a data acquisition and processing unit connected with the sensor network unit, used for collecting, synchronizing and preprocessing sensor data; an algorithm processing unit; a fusion decision unit connected with the algorithm processing unit, used for integrating multi-algorithm output results, calculating a final compensation value, and dynamically adjusting algorithm weights according to environmental conditions; and an execution control unit connected with the fusion decision unit, used for adjusting the position of a state indication mechanism according to the output of the fusion decision unit. The application realizes ultra-high precision, full environmental adaptability and intelligent self-optimization capability of switch state indication through deep fusion of a mechanical deviation correction algorithm, a nonlinear thermodynamic compensation neural network algorithm and a quantum acoustic spectrum vibration suppression algorithm.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power system protection devices, in particular to a self-adaptive small circuit breaker switch state indication system for realizing high-precision, full-environment-adaptive indication and predictive maintenance function of small circuit breaker switch state. BACKGROUND

[0002] As a key protection device in power systems, the accuracy of small circuit breaker switch state indication is crucial for system safe operation and fault diagnosis. In the prior art, there are assembly problems in the state switching process of small circuit breaker devices, which cause deviation in the switch state indication position, affecting the accurate control and reliable operation of the system. The above deviation is mainly caused by factors such as tolerance accumulation in the mechanical structure assembly process, unstable contact between the contact point and the mechanical parts during the state switching process, position deviation during the movement of the rod, and wear of the mechanical parts after long-term operation.

[0003] Traditional solutions mainly focus on mechanical structure optimization and simple single algorithm compensation, such as reducing deviation by improving rod shape and adjusting contact point position, or using linear compensation algorithm to correct the indication position. Although the above methods can improve the indication accuracy to a certain extent, they have obvious limitations: they can only deal with deviation caused by a single physical factor, and have poor adaptability to environmental factors such as temperature change, vibration, and material aging, and it is difficult to realize long-term stable and high-precision state indication.

[0004] In recent years, although some advanced algorithm technologies have appeared in the power industry, such as thermodynamic compensation algorithm based on neural network and vibration suppression algorithm based on quantum computing idea, the application of the above advanced algorithms in the field of small circuit breakers faces the following problems:

[0005] Most of the controllers deployed in the field use low-cost MCUs, which cannot support real-time calculation of neural networks. At the same time, advanced algorithms usually require multiple sensor inputs (such as temperature field distribution, multi-point vibration detection, etc.), increasing the system degree and potential failure points, which is contrary to the simple and reliable design principle of power equipment. In addition, the reliability verification of power equipment requires long-term testing, especially for algorithms that need to simulate temperature cycles and aging processes, complete verification may take several years, which does not meet the needs of rapid iteration in the industry. When multiple algorithms are simply superimposed, they often interfere with each other, such as temperature changes affecting vibration characteristics, and vibration changing temperature distribution, which is difficult to handle with simple superposition.

[0006] Therefore, there is an urgent need in the industry for a system solution that can overcome the above obstacles, effectively integrate the advantages of advanced algorithms, and adapt to the actual working conditions of power equipment. SUMMARY

[0007] The purpose of the present application is to provide a self-adaptive miniature circuit breaker switch state indication system, which realizes ultra-high precision, full environmental adaptability and intelligent self-optimization capability of switch state indication through deep fusion of mechanical deviation correction algorithm, nonlinear thermodynamic compensation neural network algorithm and quantum acoustic spectrum vibration suppression algorithm.

[0008] Unlike the traditional simple superposition method, the present application proposes a "core-shell-net" three-layer fusion architecture and algorithm cross-fusion mechanism, realizing deep mutual penetration and synergistic enhancement among algorithms. By establishing a unified multi-physical field modeling framework and a comprehensive deviation compensation model, the system can perceive and adapt to mechanical assembly errors, temperature changes and vibrations and other physical interference factors, realizing the improvement of state indication precision.

[0009] The above technical purpose of the present application is realized by the following technical scheme:

[0010] A self-adaptive miniature circuit breaker switch state indication system comprises a mechanical structure unit, a sensor network unit, a data acquisition and processing unit, an algorithm processing unit, a fusion decision unit and an execution control unit.

[0011] The mechanical structure unit comprises a vertical rod, a contact point and a state indication mechanism, the vertical rod interacts with the contact point and drives the state indication mechanism to display the switch state. The system optimizes the planar structure design of the vertical rod, controls that in the open state, even if the lower contact point continues to drive the rod to move up, it will not affect the accuracy of the indication position.

[0012] The sensor network unit comprises a position sensor, a temperature sensor network and a vibration sensor, which is used for comprehensively perceiving the system state and environmental conditions. The system adopts a shared sampling clock design to control the time synchronization of multi-sensor data, providing a basis for subsequent multi-physical quantity correlation analysis.

[0013] The data acquisition and processing unit controls the acquisition, synchronization and preprocessing of sensor data, realizes cross-validation and reliability evaluation of sensor data through data fusion technology, effectively filters abnormal data and improves system stability.

[0014] The algorithm processing unit is the core of the system, which comprises an algorithm module, a thermal compensation algorithm module and a vibration suppression algorithm module. The algorithm module performs mechanical deviation correction function, the thermal compensation algorithm module realizes nonlinear compensation of temperature influence, and the vibration suppression algorithm module processes the deviation caused by vibration. Instead of independent operation, each algorithm module realizes deep fusion through parameter sharing, feature cross and other mechanisms.

[0015] The fusion decision unit controls the output results of the integrated multi-algorithm, dynamically adjusts the weights of each algorithm through the adaptive meta-controller, handles the potential conflicts between algorithms, and generates the final compensation decision.

[0016] The execution control unit accurately adjusts the position of the state indication mechanism according to the output of the fusion decision unit, and controls the accuracy of the switch state indication. At the same time, the system also has a predictive maintenance function, which can predict possible failures based on historical data analysis and issue maintenance warnings in advance.

[0017] The system provides a unified comprehensive deviation compensation model, which integrates mechanical deviation, temperature influence and vibration interference into a multi-source unified framework:

[0018] ΔP total (t)=ΔP mech (t)+ΔP therm (t)+ΔP vib (t)+ΔP coupled (t);

[0019] Wherein, the key is to introduce the term ΔP coupled (t), which describes the coupling deviation generated by the interaction of different physical factors, and is accurately controlled through the mechanical-thermal-vibration cross-influence model, solving the problem of physical interaction that cannot be handled by simple superposition.

[0020] The present application has the following beneficial effects:

[0021] Ultra-high precision indication: the indication accuracy of the system is improved from the original ±0.2mm to ±0.02mm, and the high precision of ±0.12mm can still be maintained under extreme working conditions, providing reliable protection for the safe operation of the power system.

[0022] Full environmental adaptability: the temperature range is expanded to -40℃ to +85℃, the vibration resistance is improved to 2.0g, and it can cope with combined working conditions such as "low temperature + strong vibration", and is suitable for various harsh environments.

[0023] Self-learning optimization capability: the system has the characteristics of becoming more accurate with use, and the longer the use time, the stronger the adaptability. After 3 months, the temperature adaptation response time is shortened from 30 minutes to 15 minutes, and after 12 months, it is further shortened to 5 minutes.

[0024] Predictive maintenance function: it can identify early degradation trends and provide early warning of possible failures 72 hours in advance, reduce unplanned downtime losses, and significantly improve system availability.

[0025] Long-term stability: The calibration period is extended from the traditional 6 months to 36 months, and the accuracy retention rate after 5 years reaches 66.7%, which is much higher than the traditional system of 33.3%, and greatly reduces the maintenance cost.

[0026] The application realizes deep cross fusion between algorithms, brings higher performance improvement to the small circuit breaker switch state indication system, and provides a new application prospect for the intelligentization of power equipment. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 It is a structural schematic diagram of the adaptive small circuit breaker;

[0028] Figure 2 It is Figure 1 It is a schematic diagram of the internal structure after removing the side plate of the circuit breaker shell;

[0029] Figure 3 It is a front view of the adaptive small circuit breaker switch state indication system;

[0030] Figure 4 It is a schematic diagram of the "core-shell-net" three-layer fusion architecture of the system of the application;

[0031] Figure 5 It is an algorithm processing flowchart of the system of the application;

[0032] Figure 6 It is a schematic diagram of the multi-algorithm cross fusion mechanism of the system of the application;

[0033] Figure 7 It is a schematic diagram of the multi-physical field unified modeling framework of the system of the application;

[0034] Figure 8 It is a performance comparison curve diagram of the system of the application under different temperature conditions;

[0035] Figure 9 It is a performance comparison curve diagram of the system of the application under different vibration conditions. DETAILED DESCRIPTION

[0036] The application will be further described in detail below in combination with the drawings and specific embodiments, but the embodiments of the application are not limited thereto.

[0037] The adaptive small circuit breaker switch state indication system of the application comprises a mechanical structure unit 1, a sensor network unit 2, a data acquisition and processing unit 3, an algorithm processing unit 4, a fusion decision unit 5 and an execution control unit 6.

[0038] As Figure 1 , Figure 2 , Figure 3As shown, mechanical structure unit 1 includes a vertical rod 11, a contact point 12, and a status indicating mechanism 13. The vertical rod 11 is made of high-strength alloy material, and its length is typically between 10-50 mm, with the appropriate length selected according to the different models of circuit breakers. Its characteristic is that it has a large planar margin, ensuring that even if the rod continues to move upward after the contact is released in the open state, it will not cause deviation in the switch indication position. The contact point 12 is typically made of wear-resistant alloy material and contacts the vertical rod 11, forming a force transmission path. The status indicating mechanism 13 includes an indicator rod and an indicator mark, used to display the open / closed status of the switch.

[0039] Sensor network unit 2 includes position sensor 21, temperature sensor network 22, and vibration sensor 23. Position sensor 21 employs a high-precision linear displacement sensor with an accuracy of ±0.01mm, used to measure switch positions and status indication positions. Temperature sensor network 22 includes multiple distributed temperature sensors arranged in the environment and at component locations, with a measurement range of -45℃ to +90℃ and an accuracy of ±0.5℃. Vibration sensor 23 employs a triaxial accelerometer with a measurement range of 0-5g and a sampling frequency of 1kHz. All sensors share the same sampling clock, controlling data time synchronization and providing a basis for subsequent correlation analysis of multiple physical quantities.

[0040] The data acquisition and processing unit 3 employs a high-performance microcontroller equipped with a multi-channel ADC converter, a sampling rate of no less than 10kHz, and a resolution of no less than 12 bits. This unit controls the acquisition, synchronization, and preprocessing of sensor data, including functions such as data filtering, anomaly detection, and preliminary fusion, and controls subsequent algorithm processing to obtain high-quality input data.

[0041] Algorithm processing unit 4 is the core of the system, employing methods such as... Figure 4 The "core-shell-network" three-layer fusion architecture shown includes a core algorithm module 41, a thermal compensation algorithm module 42, and a vibration suppression algorithm module 43. The core algorithm module 41 is located in the core layer and performs mechanical deviation correction; the thermal compensation algorithm module 42 and the vibration suppression algorithm module 43 are located in the shell layer and handle the effects of temperature and vibration, respectively; the three modules achieve deep fusion through parameter sharing and feature crossing in the network layer.

[0042] The fusion decision unit 5 controls and integrates the output results of multiple algorithms, including a meta-controller 51 and a conflict resolution module 52. The meta-controller 51 dynamically adjusts the weights of each algorithm based on environmental conditions, and the conflict resolution module 52 handles potential conflicts between algorithms to jointly generate a compensation decision.

[0043] The execution control unit 6 is connected to the fusion decision unit 5 and includes a drive circuit 61 and an execution mechanism 62. It controls the precise adjustment of the position of the status indicator mechanism according to the output of the fusion decision unit, thereby controlling the accuracy of the switch status indication.

[0044] The core of this invention lies in the cross-fusion mechanism between algorithms, such as Figure 5 , Figure 6 As shown, unlike the traditional simple superposition method, this system achieves deep cross-integration between algorithm steps and formulas.

[0045] 1. Based on the algorithm module

[0046] The algorithm module executes three core algorithms: the switch position deviation model, the lever motion model in the open state, and the mechanical deviation correction algorithm.

[0047] The switch position deviation is represented by the model as follows:

[0048] ΔP=f(L,θ,F c );

[0049] Wherein, ΔP represents the deviation of the switch position, in millimeters (mm), indicating the difference between the actual indicated position and the ideal position, typically ranging from 0 to 2 mm; L represents the length of the vertical rod, in millimeters (mm), typically within the range of 10-50 mm; θ represents the angular deviation when the vertical rod contacts the contact plane, in degrees (°), with an ideal value of 0°, and in practice usually controlled within ±2°; F c This indicates the magnitude of the force applied at the contact point, measured in Newtons (N), with typical values ​​ranging from 5 to 20 N.

[0050] The function f represents the nonlinear relationship between the deviation and the three parameters, obtained by fitting experimental data. In practical applications, it is usually represented by a piecewise polynomial. For example, for L in the range of 20-30 mm, it can be expressed as:

[0051] f(L,θ,F c = 0.05*L + 0.1*θ + 0.01*F c +0.002*L*θ;

[0052] The motion model of the lever under the open state is represented as follows:

[0053] F c =k·(x-x0);

[0054] Among them, F c The force at the contact point is expressed in Newtons (N), ranging from 5 to 20 N; k represents the elastic modulus of the rod, expressed in Newtons per millimeter (N / mm), ranging from 2 to 10 N / mm, reflecting the material and structural characteristics of the rod; x represents the current displacement of the rod, expressed in millimeters (mm); x0 represents the initial contact position, expressed in millimeters (mm).

[0055] The above model describes the relationship between force and displacement in a linear elastic system, showing the variation of contact force with member displacement during the opening process. The member elastic coefficient k is usually optimized through material selection and heat treatment processes to obtain stable elastic properties.

[0056] The mechanical deviation correction algorithm is expressed as follows:

[0057] R = ΔP - α·(L - L0);

[0058] Where R represents the corrected switch indication position in millimeters (mm); α represents the adjustment coefficient, which is dimensionless and typically ranges from 0.05 to 0.5; and L0 represents the ideal length of the rod in millimeters (mm).

[0059] The algorithm described above dynamically compensates for deviations caused by changes in member length by introducing an adjustment coefficient α, thus achieving the correction of mechanical deviations. While α is typically a fixed value, in this invention, α is extended to a dynamic parameter, forming a parameter-sharing relationship with the thermal compensation and vibration suppression algorithms.

[0060] 2. Thermal Compensation Algorithm Module

[0061] The thermal compensation algorithm module executes a nonlinear thermodynamic compensation neural network algorithm, including a temperature-deviation nonlinear mapping model and a material thermal expansion nonlinear coefficient model.

[0062] The temperature-deviation nonlinear mapping model is expressed as:

[0063] ΔP T (t)=Σ[w i ·σ(v i ·T(t)+b i )]+γ·dT(t) / dt;

[0064] Where, ΔP T (t) represents the time-varying deviation caused by temperature, in millimeters (mm); T(t) represents the current temperature field distribution, in degrees Celsius (°C), and is a temperature vector including ambient temperature and component temperature; w i ,v i ,b i The weights and bias parameters of the neural network are obtained through training, w i The typical range is -1.0 to 1.0, v i The range is from -0.5 to 0.5, b iThe range is -2.0 to 2.0; σ represents the activation function, usually ReLU or tanh function; γ represents the temperature change rate coefficient, in millimeters / (degrees Celsius·second), with a typical value of 0.001-0.01; dT(t) / dt represents the temperature change rate, in degrees Celsius / second (°C / s).

[0065] The above model utilizes a deep neural network to establish a nonlinear mapping relationship between temperature changes and state indication deviations, overcoming the limitations of traditional linear thermal compensation models. The model consists of two parts: the first part Σ[w i ·σ(v i ·T(t)+b i The first part describes the static effect of temperature, while the second part, γ·dT(t) / dt, describes the dynamic effect of the rate of temperature change, which can effectively handle transient deviations when the temperature changes rapidly.

[0066] The nonlinear coefficient model of material thermal expansion is expressed as:

[0067] α T (T) = α0 + α1·T + α2·T 2 +α3·exp(T / T c );

[0068] Where, α T (T) represents the temperature-dependent coefficient of thermal expansion, in units of 1 degree Celsius (1 / ℃); α0, α1, α2, α3 represent the fitting coefficients, with typical values ​​of 1.0 × 10^-5, 1.0 × 10^-7, 1.0 × 10^-9, and 1.0 × 10^-6, respectively; T represents the current temperature, in units of degrees Celsius (℃); T c It represents the critical temperature coefficient of a material, with the unit being degrees Celsius (°C), and a typical value of 100-200°C.

[0069] The above model considers the nonlinear thermal expansion characteristics of the material in different temperature ranges, especially through the exponential term exp(T / T) c This model captures the nonlinear changes of materials near their critical temperature, improving compensation accuracy under extreme temperature conditions. Unlike traditional linear thermal expansion coefficient models, the aforementioned nonlinear model can accurately describe the thermal response characteristics of materials over a wide temperature range.

[0070] Based on the two models mentioned above, the thermal compensation algorithm also calculates the change in rod length caused by temperature:

[0071] δ L (T)=L0·∫[α T [(T)·dT]+σ th (T) / E(T)·L0;

[0072] Where, δ L(T) represents the change in member length due to temperature, in millimeters (mm); L0 represents the initial member length, in millimeters (mm); σ th (T) represents thermal stress, in megapascals (MPa); E(T) represents temperature-dependent Young's modulus, in gigapascals (GPa).

[0073] The above calculation results are not only used for thermal compensation, but also passed to the algorithm module through the parameter sharing mechanism to dynamically adjust the adjustment coefficient α, thereby realizing the cross-fusion between algorithms.

[0074] 3. Vibration Suppression Algorithm Module

[0075] The vibration suppression algorithm module executes the quantum acoustic spectrum vibration suppression algorithm, the core of which is the vibration-deviation mapping matrix model:

[0076] ΔP V =M·V+V^T·N·V;

[0077] Where, ΔP V The vector represents the deviation caused by vibration, in millimeters (mm); M represents the linear mapping matrix with dimensions n×m, where n is the dimension of the deviation vector and m is the dimension of the vibration feature vector; N represents the second-order nonlinear mapping tensor with dimensions n×m×m; V represents the vibration feature vector, containing vibration frequency and amplitude information; V^T represents the transpose of V.

[0078] The model described above provides a mapping relationship between vibration characteristics and switch state indication deviations, comprising two parts: linear influence (M·V) and nonlinear coupling influence (V^T·N·V). The linear mapping matrix M describes the direct contribution of vibrations at each frequency to the deviation, while the nonlinear mapping tensor N captures the interactive influence between different vibration frequencies, an influence that is difficult to handle by traditional vibration models.

[0079] The vibration feature vector V is obtained by performing time-frequency analysis on the vibration signal:

[0080] V(f,t)=Σ[A i (t)·exp(j·2πf i t)·exp(-ξ i ·t)];

[0081] Where V(f,t) represents the time-frequency representation of the vibration; A i (t) represents the amplitude of the i-th vibration component, in micrometers (μm); f i ξ represents the frequency of the i-th vibration component, in Hertz (Hz); i represents the damping coefficient, with units of 1 / second (s^-1); j represents the imaginary unit.

[0082] The algorithm described above can identify and separate three vibration sources in real time: environmental vibration, operational vibration, and natural oscillation, and specifically suppress the impact of each type of vibration on the switch state indication. By processing vibration modes using the superposition principle in quantum computing, the system can calculate the state probability distribution under different vibration combinations:

[0083] P(s|V)=|<s|U(V)|s0> | 2 ;

[0084] Where P(s|V) represents the probability that the system is in state s given vibration V; |s0> represents the initial state of the system; U(V) represents the vibration-related evolution operator; |<s|U(V)|s0> | 2 This represents the probability of superposition of states in quantum mechanics.

[0085] The aforementioned model enables the system to predict possible states under vibration conditions, significantly improving its resistance to vibration disturbances. Simultaneously, the algorithm calculates the impact of vibration on the elastic coefficients of the members and transmits the results to the underlying algorithm module, demonstrating the cross-influence between algorithms.

[0086] 4. Multi-algorithm collaborative fusion

[0087] like Figure 7 As shown, the core technology of this invention lies in the synergistic fusion of multiple algorithms, rather than simple superposition. The system provides a unified comprehensive deviation compensation model:

[0088] ΔP total (t)=ΔP mech (t)+ΔP therm (t)+ΔP vib (t)+ΔP coupled (t);

[0089] Where, ΔP total (t) represents the total deviation, in millimeters (mm); ΔP mech (t) represents the deviation caused by mechanical assembly, which comes from the algorithm module; ΔP therm (t) represents the deviation caused by temperature change, obtained from the self-heating compensation algorithm module; ΔP vib (t) represents the deviation caused by vibration, obtained through the vibration suppression algorithm module; ΔP coupled (t) represents the interactive bias caused by the coupling of multiple factors.

[0090] ΔP was introduced. coupled Term (t) describes the coupling deviation caused by the interaction of different physical factors, a problem that cannot be solved by simple superposition. The system precisely controls the above coupling effects through a mechanical-thermal-vibration cross-influence model:

[0091] ΔPcoupled (t)=λ TM ·f TM (T,L,k)+λ VM ·f VM (V,F c ,x)+λ TV ·f TV (T,V);

[0092] Where, λ TM ,λ VM ,λ TV denoted by f, the coupling coefficient is obtained through adaptive learning and typically ranges from 0.1 to 0.5. TM (T,L,k) represents the effect function of temperature on mechanical properties; f VM (V,F c f(x) represents the effect function of vibration on contact force and displacement; TV (T,V) represents the interaction function between temperature and vibration.

[0093] The three cross-influence functions are as follows:

[0094] f TM (T,L,k)=[k(T) / k0-1]·(x-x0)+α T (T)·(T-T0)·L;

[0095] This represents the effect of temperature on the elastic modulus and length of the bar, where k(T) is the temperature-dependent elastic modulus and k0 is the elastic modulus at the reference temperature.

[0096] f VM (V,F c ,x)=μ·∫V(f,t)·H F (f)·d f ·F c +∫V(f,t)·H x (f)·d f ;

[0097] This represents the effect of vibration on contact force and displacement, where μ is the coefficient of friction, and H... F (f) and H x (f) is the frequency-dependent transfer function;

[0098] f TV (T,V)=ξ(T)·∫V(f,T)·S(f,T)·d f ;

[0099] This represents the interaction between temperature and vibration, where ξ(T) is the temperature-dependent coupling coefficient and S(f,T) is the temperature-dependent frequency sensitivity function.

[0100] The aforementioned cross-influence function demonstrates the interaction between physical fields, the change in material elastic modulus caused by temperature changes, the thermal expansion effect, and the influence of temperature on the vibration transfer function, thus achieving accurate modeling of the multi-physics coupling effect.

[0101] Based on the above model, the fusion decision unit executes the final fusion decision compensation algorithm:

[0102] R Fusion (t)=ΔP total (t)-[w M ·α·(L(t)-L0)+w T ·(ΔP T (t)+β·δ L (T(t)))+w V ·(Σ[w i ·ΔP Vi ]+η·∫V(f,t)·S(f)·d f )+w C ·ΔP coupled (t)];

[0103] Among them, R Fusion (t) indicates the final switch indicator position after fusion compensation; w M ,w T ,w V ,w C The values ​​in square brackets represent the dynamic weights of each compensation term, determined by environmental conditions; the terms in square brackets are the original mechanical compensation term, temperature compensation term, vibration compensation term, and coupling compensation term, respectively.

[0104] The weighting factors are dynamically adjusted through an adaptive mechanism:

[0105] w i (t)=softmax(σ i ·E i (t));

[0106] Where, σ i E represents the reliability coefficient of the i-th algorithm, typically ranging from 0.5 to 2.0; i (t) represents the current environment fitness score of the i-th algorithm, reflecting the applicability of the algorithm in the current environment.

[0107] The aforementioned dynamic weighting mechanism controls the system to automatically select the optimal algorithm combination based on environmental conditions: when the temperature changes drastically but the vibration is small, the system will increase the weight of the thermal compensation algorithm; when the vibration is strong but the temperature is stable, the weight of the vibration suppression algorithm will be increased; and under extreme combined working conditions, the role of the coupling compensation term will be fully utilized.

[0108] The algorithmic cross-fusion of this invention is not only reflected at the model level, but also in multiple aspects such as parameter sharing, feature cross-fertilization, prediction feedback, and learning signal sharing.

[0109] 1. Parameter sharing and dynamic adjustment

[0110] The three algorithm modules do not operate in isolation, but rather achieve cross-influence by sharing key physical parameters:

[0111] The temperature-dependent elastic coefficient k(T) calculated by the thermal compensation algorithm is passed to the algorithm module to dynamically adjust the k value in the rod motion model; the vibration characteristics detected by the vibration suppression algorithm affect the friction coefficient μ, and thus the contact force F. c The calculations are performed based on the member length L and displacement x of the algorithm module, which are simultaneously shared with the thermal compensation and vibration suppression algorithms as important inputs. This parameter sharing ensures that the algorithms no longer operate based on static or independent parameters, but rather on dynamically updated globally consistent parameters, thus controlling the synergy between the algorithms.

[0112] 2. Feature Cross Enhancement

[0113] Temperature and vibration features are combined to form enhanced features; that is, the product of temperature gradient and vibration amplitude is used to predict possible nonlinear responses. Composite features are formed by combining mechanical and environmental parameters, such as the combination of contact angle and temperature, to predict thermal stress distribution. Feature extraction networks are used to reduce the dimensionality and enhance the raw sensor data, extracting more informative features. These feature interactions enable the system to capture physical phenomena that a single feature cannot reflect, improving the model's expressive power.

[0114] 3. Predictive feedback mechanism

[0115] The temperature prediction of the thermal compensation algorithm directly affects the vibration characteristic analysis of the vibration suppression algorithm. For example, if a temperature rise is predicted, the vibration transfer function will be adjusted in advance. The vibration analysis results are fed back to the thermal compensation algorithm to adjust the convective heat transfer coefficient in the thermal model. Based on the changes in contact state detected by the algorithm, the parameters of the thermal compensation and vibration suppression algorithms will be re-evaluated. The above predictive feedback enables the system to respond to changes in physical state in advance, achieve predictive adjustments, and reduce response lag.

[0116] 4. Learning signal sharing

[0117] The three algorithm modules share the same error signal for collaborative optimization:

[0118] e(t) = R m (t)-R Fusion (t);

[0119] Where e(t) is the error between the measured position and the predicted position, and this error signal is used to update the parameters of the three algorithm modules:

[0120] Adjust the adjustment coefficient α according to the algorithm module:

[0121] α(t+1)=α(t)+η α ·e(t)·(L-L0);

[0122] The heat compensation algorithm module adjusts the neural network weights:

[0123] w i (t+1)=w i (t)+η w ·e(t)· ;

[0124] The vibration suppression algorithm module adjusts the mapping matrix:

[0125] M(t+1)=M(t)+η M ·e(t)·V^T;

[0126] By sharing learning signals, the three algorithms form a unified learning system, controlling the consistency and coordination of the optimization direction.

[0127] 5. Adaptive Meta Controller

[0128] The system introduces a meta-controller to coordinate the control algorithms:

[0129] {w M ,w T ,w V ,w C}=MetaController(E M E T E V ,ΔR history ,dE / dt);

[0130] The meta-controller determines the optimal collaboration mode among learning algorithms, thus forming algorithm management. By analyzing the environmental adaptability, historical performance, and error change trends of each algorithm, it dynamically adjusts the optimal fusion strategy to achieve the system's adaptive capability.

[0131] The meta-controller also handles algorithm conflicts. When there are significant differences in algorithm outputs, it resolves conflicts through methods such as confidence assessment, median filtering, and augmented sampling, thereby improving the consistency and reliability of system decision-making.

[0132] The complete workflow of the system of this invention is as follows:

[0133] 1. System Initialization

[0134] When the system starts, it first performs initial configuration: loads mechanical structure parameters: ideal member length L0, elastic coefficient k, and initial value of adjustment coefficient α; initializes the NTCNN pre-trained model and thermal expansion coefficient parameters; prepares the QASVA vibration feature library and model parameters; performs sensor self-test and calibration, and establishes sensor drift baseline; evaluates the current environmental conditions, and selects the initial algorithm fusion mode.

[0135] 2. Data Acquisition and Preprocessing

[0136] The system continuously collects data from multimodal sensors: position sensors measure switch positions and status indicator positions; temperature sensor networks monitor ambient temperature and component temperature distribution; vibration sensors monitor system vibration status; a synchronous triggering mechanism controls data time alignment; and the raw data is filtered, denoised, and anomaly detected.

[0137] 3. Parallel processing of multiple algorithms

[0138] The three algorithm modules execute their respective computational tasks in parallel:

[0139] (1) Based on the algorithm module:

[0140] Calculate the switch position deviation ΔP = f(L, θ, F) c According to the linkage motion model F c =k·(x-x0) Analyze the contact state; preliminarily calculate the mechanical compensation amount α·(L-L0).

[0141] (2) Thermal compensation algorithm module: calculates temperature deviation ΔP through neural network. T (t); Calculate the change in rod length δ based on the thermal expansion model. L (T); Evaluate the effect of temperature on material properties and update the k(T) value.

[0142] (3) Vibration suppression algorithm module: Analyze the vibration spectrum characteristics, distinguish between environmental vibration, operational vibration and natural oscillation; calculate the deviation ΔP caused by vibration. V ;Assess the impact of vibration on the contact state.

[0143] 4. Cross-impact assessment

[0144] The system calculates the cross-influence between physical factors: the effect of temperature on mechanical properties. TM (T,L,k); Effect of vibration on contact force and displacement f VM (V,F c ,x); The interaction between temperature and vibration f TV (T,V); Comprehensive evaluation of coupling deviation ΔP coupled (t).

[0145] 5. Fusion Decision Generation

[0146] The fusion decision unit integrates the results of multiple algorithms to generate the final decision.

[0147] The fitness E of each algorithm is evaluated based on the current environmental conditions. i (t); The algorithm weights w are dynamically adjusted by the meta-controller. i (t); Calculate the total deviation ΔP total (t) and fusion compensation value R Fusion (t); Detect whether there is a conflict in the algorithm output, and start the conflict resolution mechanism if there is a conflict.

[0148] 6. Execution Control and Feedback

[0149] The execution control unit performs compensation adjustments:

[0150] The drive actuator adjusts the position of the status indicator mechanism; the actual position R is measured by a position sensor. m (t); Calculation error e(t) = R m (t)-R Fusion (t); The error is fed back to each algorithm module for parameter updates.

[0151] 7. Adaptive Learning and Optimization

[0152] The system performs adaptive learning based on feedback error:

[0153] Update the adjustment coefficient α based on the algorithm; adjust the neural network weights of the NTCNN model; optimize the mapping matrix and tensor of QASVA; update the weight allocation strategy of the meta-controller; save the learning results for future initialization.

[0154] 8. Condition monitoring and predictive maintenance

[0155] The system continuously monitors its status and performs health assessments:

[0156] Analyze parameter change trends to detect potential signs of degradation; compare current performance with historical baselines to assess system health; predict potential failures and issue maintenance warnings 72 hours in advance; generate performance reports and provide maintenance recommendations.

[0157] Based on the above discussion, the system of this invention achieves high-precision switching status indication through the collaborative fusion of multiple algorithms under environmental conditions. The system can accurately capture mechanical factors, temperature changes, vibration effects, and their interactions, and integrates the results of multiple algorithms through an adaptive weighting mechanism to achieve precise compensation control. This deep fusion approach far surpasses the capabilities of traditional technologies and simple algorithm superposition, providing a reliable guarantee for the safe and stable operation of power systems.

[0158] To better understand the technical implementation process and synergistic integration effect of the present invention, this application provides the following detailed calculation process to demonstrate the operating mechanism and performance of the system under different working conditions.

[0159] 1. System parameters and initial settings under extreme low temperature conditions

[0160] Suppose a miniature circuit breaker is installed in a substation in a high-altitude, cold region, where the operating environment is harsh. The system has been running for 12 months, and its current operating parameters are as follows:

[0161] (1) Environment and operating parameters:

[0162] Ambient temperature (T) env The temperature is -35℃, and there is a rapid temperature change, with a temperature change rate (dT / dt) of 0.5℃ / minute.

[0163] Vibration environment: Mixed vibration of 50Hz fundamental frequency and 120Hz high frequency, with amplitudes of 0.8g and 0.3g respectively;

[0164] The system's cumulative running time (t) = 8760 hours;

[0165] Total number of operations (N) = 2880;

[0166] Last calibration date: 6 months ago.

[0167] (2) Mechanical structure parameters:

[0168] The length of the vertical member (L) is 24.85 mm (affected by temperature shrinkage).

[0169] Ideal member length (L0) = 25.00 mm;

[0170] Angular deviation (θ) = 0.8°;

[0171] Contact force (F) c =15.2N;

[0172] Initial contact position (x0) = 10.00 mm;

[0173] Current displacement (x) = 13.20 mm;

[0174] The elastic modulus (k0) at the reference temperature is 5.0 N / mm.

[0175] (3) Sensor network readings:

[0176] Position sensor: Status indication position = 1.78mm;

[0177] Temperature sensor network: [-35℃, -32℃, -30℃] (ambient, rod surface, contact point);

[0178] Vibration sensor: x-axis 0.75g@50Hz, y-axis 0.32g@50Hz, z-axis 0.25g@120Hz;

[0179] (4) Algorithm parameters:

[0180] ①NTCNN neural network parameters: weights w=[0.15,0.22,0.18,-0.12]; input transformation coefficients v=[0.25,0.15,0.30,-0.20]; bias b=[0.8,1.2,-0.5,0.3]; temperature change rate coefficient γ=0.012mm / (℃·min).

[0181] Thermal expansion coefficient parameters: α0 = 1.2 × 10⁻⁵ / ℃; α1 = 2.5 × 10⁻⁷ / ℃²; α2 = 1.8 × 10⁻⁹ / ℃³; α3 = 9.5 × 10⁻⁶; T c =180℃.

[0182] ②QASVA algorithm parameters:

[0183] The linear mapping matrix M = [[0.015, 0.0003, 0.18], [0.008, 0.0005, 0.12]]; the nonlinear mapping coefficient N = 0.003; and the damping coefficient ξ = [0.06, 0.08].

[0184] ③ Cross-influence parameter: λ TM =0.35 (temperature-mechanical coupling coefficient); λ VM =0.25 (vibration-mechanical coupling coefficient); λ TV =0.15 (temperature-vibration coupling coefficient); friction coefficient μ=0.22; thermal expansion compensation coefficient β=0.85.

[0185] 2. Multi-algorithm collaborative fusion computation process

[0186] (1) Calculate based on the algorithm module

[0187] Step 1: Calculate the basis value for switch position deviation

[0188] Application of switch position deviation based on model:

[0189] ΔP mech =f(L,θ,F c )

[0190] =0.05*L+0.1*θ+0.01*F c +0.002*L*θ

[0191] =0.05*24.85+0.1*0.8+0.01*15.2+0.002*24.85*0.8=1.2425+0.08+0.152+0.03976=1.51426mm.

[0192] Step 2: Verify the lever motion model under the open state

[0193] In low-temperature environments, the temperature-corrected elastic modulus needs to be calculated first:

[0194] β k =-0.0015 / ℃ (Temperature coefficient of elastic modulus of material at low temperature)

[0195] k(-35℃) = k0·[1+β] k ·(-35-20)]

[0196] =5.0·[1+(-0.0015)·(-55)]=5.0·[1+0.0825]=5.0·1.0825=5.4125N / mm.

[0197] Based on the corrected elastic coefficient, the relationship between contact force and displacement is verified:

[0198] Theoretical contact force = k(-35℃)·(x-x0)

[0199] =5.4125·(13.20-10.00)

[0200] =5.4125·3.20=17.32N.

[0201] The actual measured value was 15.2 N, with a difference of 2.12 N (approximately 12.2%).

[0202] The significant difference indicates the presence of additional factors in the low-temperature environment, such as icing at the contact surface or changes in the material's stress state.

[0203] Step 3: Perform mechanical deviation correction calculation

[0204] Using the system's current adjustment coefficient α=0.25 (optimized through adaptive learning):

[0205] L-L0=24.85-25.00=-0.15mm;

[0206] α·(L-L0)=0.25·(-0.15)=-0.0375mm;

[0207] R mech =ΔP mech -α·(L-L0)=1.51426-(-0.0375)=1.55176mm.

[0208] (2) Calculation of thermal compensation algorithm module

[0209] Step 1: Calculate the temperature-deviation nonlinear mapping

[0210] Temperature input is processed using the ReLU activation function (σ(x)=max(0,x)):

[0211] T(t) = [-35, -32, -30] (ambient temperature, rod surface temperature, contact point temperature).

[0212] Input transformation:

[0213] v1·T+b1=0.25·(-35)+0.15·(-32)+0.30·(-30)+0.8=-8.75-4.8-9.0+0.8=-21.75 (negative value, ReLU output is 0).

[0214] v2·T+b2=0.15·(-35)+0.15·(-32)+0.30·(-30)+1.2=-5.25-4.8-9.0+1.2=-17.85 (negative value, ReLU output is 0).

[0215] v3·T+b3=0.25·(-35)+0.15·(-32)+0.30·(-30)+(-0.5)=-8.75-4.8-9.0-0.5=-23.05 (negative value, ReLU output is 0).

[0216] v4·T+b4=-0.20·(-35)+0.15·(-32)+0.30·(-30)+0.3=7.0-4.8-9.0+0.3=-6.5 (negative value, ReLU output is 0).

[0217] Static temperature deviation section:

[0218] Σ[w i ·σ(v i ·T+b i )]=0.15·0+0.22·0+0.18·0+(-0.12)·0=0mm.

[0219] Since extreme low temperatures suppress all neurons (ReLU output is 0), the static temperature deviation is calculated to be 0. At this point, the dynamic temperature change component comes into play:

[0220] Temperature change rate section:

[0221] γ·dT(t) / dt=0.012·0.5=0.006mm / min.

[0222] Total temperature deviation:

[0223] ΔP T (t) = Static deviation + Dynamic deviation = 0 + 0.006 = 0.006 mm / min = 0.0001 mm / s.

[0224] Step 2: Calculate the nonlinear coefficient of thermal expansion of the material.

[0225] Calculation of the coefficient of thermal expansion of materials at extremely low temperatures:

[0226] T = -35℃;

[0227] α T (T) = α0 + α1·T + α2·T 2 +α3·exp(T / T c )

[0228] =1.2×10^-5+2.5×10^-7·(-35)+1.8×10^-9·(-35) 2 +9.5×10^-6·exp((-35) / 180)

[0229] =1.2×10^-5-8.75×10^-6+2.205×10^-6+9.5×10^-6·0.8236

[0230] =1.2×10^-5-8.75×10^-6+2.205×10^-6+7.8242×10^-6=1.3279×10^-5 / ℃.

[0231] Step 3: Calculate the change in rod length caused by temperature

[0232] Reference temperature T0 = 20℃;

[0233] ΔT = -35 - 20 = -55℃;

[0234] δ L (T)≈L0·α T (T)·ΔT=25.00·1.3279×10^-5·(-55)=-0.01828mm.

[0235] (3) Vibration suppression algorithm module calculation

[0236] Step 1: Vibration Feature Vector Extraction and Time-Frequency Analysis

[0237] Main vibration components:

[0238] f1=50Hz, A 1x =0.75g, A 1y =0.32g, A1z =0;

[0239] f2=120Hz, A 2x =0,A 2y =0,A 2z =0.25g.

[0240] Vibration eigenvector V:

[0241] V=[0.75,0.32,0,50,0,0,0.25,120].

[0242] Simplified to a two-dimensional eigenvector (composite amplitude and frequency):

[0243] V1=[0.82,50] (low-frequency components, expressed as sqrt(0.75)) 2 +0.32 2 )Calculate the composite amplitude.

[0244] V2=[0.25,120] (high frequency components).

[0245] Step 2: Calculate the deviation caused by vibration

[0246] Calculate the linear effect using the linear mapping matrix:

[0247] M·V1=[0.015,0.0003]·[0.82,50]=0.015·0.82+0.0003·50=0.0123+0.015=0.0273mm.

[0248] M·V2=[0.008,0.0005]·[0.25,120]=0.008·0.25+0.0005·120=0.002+0.06=0.062mm.

[0249] Calculate the effects of nonlinear coupling:

[0250] V1^T·N·V1≈N·||V1|| 2 =0.003·0.82 2 =0.003·0.6724=0.00202mm.

[0251] V2^T·N·V2≈N·||V2|| 2 =0.003·0.25 2 =0.003·0.0625=0.00019mm.

[0252] Total vibration deviation:

[0253] ΔP V=(M·V1+V1^T·N·V1)+(M·V2+V2^T·N·V2)

[0254] =(0.0273+0.00202)+(0.062+0.00019)=0.02932+0.06219=0.09151mm.

[0255] (4) Cross-impact assessment

[0256] Step 1: Calculate the effect of temperature on mechanical properties

[0257] f TM (T,L,k)=[k(T) / k0-1]·(x-x0)+α T (T)·(T-T0)·L=[5.4125 / 5.0-1]·3.20+1.3279×10^-5·(-55)·24.85=0.0825·3.20+1.3279×10^-5·(-55)·24.85=0.264-0.01814=0.24586mm.

[0258] Step 2: Calculate the effect of vibration on contact force and displacement

[0259] H F (50Hz) = 1.5 (the vibration transfer function value increases at low temperatures);

[0260] H F (120Hz) = 2.2 (high-frequency vibration transmission amplification factor);

[0261] f VM (V,F c ,x)=μ·∫V(f,t)·H F (f)·d f ·F c +∫V(f,t)·H x (f)·d f ≈μ·[V1·H F (50Hz)+V2·H F (120Hz)]·F c +V1·H x (50Hz)+V2·H x (120Hz)≈0.22·[0.82·1.5+0.25·2.2]·15.2+0.82·0.03+0.25·0.05=0.22·[1.23+0.55]·15.2+ 0.0246+0.0125=0.22·1.78·15.2+0.0371=0.22·27.056+0.0371=5.95232+0.0371=5.98942mm.

[0262] This value is higher than the calculation result at room temperature, indicating that the impact of vibration on contact force is significantly enhanced under extreme low temperature conditions.

[0263] Step 3: Calculate the interaction between temperature and vibration

[0264] ξ(-35℃)=1.2 (coupling coefficient increases at low temperatures);

[0265] S(50Hz, -35℃) = 1.8 (Frequency sensitivity increases at low temperatures);

[0266] S(120Hz, -35℃)=2.5 (High frequencies are more sensitive at low temperatures).

[0267] f TV (T,V)≈ξ(T)·[V1·S(f1,T)+V2·S(f2,T)]=1.2·[0.82·1.8+0.25·2.5]=1.2·[1.476+0.625]=1.2·2.101=2.5212mm.

[0268] Step 4: Calculate the total interaction bias caused by the coupling of multiple factors.

[0269] ΔP coupled (t)=

[0270] λ TM ·f TM (T,L,k)+λ VM ·f VM (V,F c ,x)+λ TV ·f TV (T,V)=

[0271] 0.35·0.24586+0.25·5.98942+0.15·2.5212=0.08605+1.49736+0.37818=1.96159mm.

[0272] (5) Calculation of overall deviation

[0273] Calculate the total deviation:

[0274] ΔP total (t)=

[0275] ΔP mech (t)+ΔP therm (t)+ΔP vib (t)+ΔP coupled (t)=1.51426+0.0001+0.09151+1.96159=3.56746mm.

[0276] (6) Fusion Decision Generation

[0277] Step 1: Environmental Condition Assessment and Weight Calculation

[0278] The system evaluates the fitness of each algorithm under the current operating conditions based on the current extreme environmental conditions and sensor readings:

[0279] Environmental adaptability score:

[0280] E M =0.4 (The fitness of mechanical models decreases at extremely low temperatures);

[0281] E T =0.9 (significant temperature changes indicate high adaptability of the thermal compensation algorithm);

[0282] E V =0.7 (moderate vibration intensity);

[0283] E C =0.95 (The coupling effect is significantly enhanced under extreme conditions).

[0284] Credibility coefficient:

[0285] σ M =0.9,σ T =1.5,σ V =1.2,σ C =1.8.

[0286] Calculation of softmax function weights:

[0287] Denominator = exp(0.9·0.4) ​​+ exp(1.5·0.9) + exp(1.2·0.7) + exp(1.8·0.95)

[0288] =exp(0.36)+exp(1.35)+exp(0.84)+exp(1.71)=1.433+3.857+2.316+5.529=13.135.

[0289] w M =1.433 / 13.135=0.109;

[0290] w T =3.857 / 13.135=0.294;

[0291] w V =2.316 / 13.135=0.176;

[0292] w C =5.529 / 13.135=0.421.

[0293] The weight allocation is clearly biased towards the coupling compensation term, and the system identifies that the cross-effect is significantly enhanced under extreme environments.

[0294] Step 2: Fusion Compensation Calculation

[0295] R Fusion (t)=ΔP total (t)-[w M ·α·(L-L0)+w T ·(ΔP T (t)+β·δ L (T(t)))+w V ·(ΔP V )+w C ·ΔP coupled (t)].

[0296] Compensation Calculation:

[0297] w M ·α·(L-L0)=0.109·0.25·(-0.15)=-0.00409mm;

[0298] w T ·(ΔP T (t)+β·δ L (T(t)))=0.294·(0.0001+0.85·(-0.01828))=0.294·(-0.01554)=-0.00457mm;

[0299] w V ·ΔP V =0.176·0.09151=0.01611mm;

[0300] w C ·ΔP coupled (t)=0.421·1.96159=0.82583mm.

[0301] Total compensation = -0.00409 - 0.00457 + 0.01611 + 0.82583 = 0.83328 mm.

[0302] R Fusion (t)=3.56746-0.83328=2.73418mm.

[0303] (7) Execution control and feedback

[0304] The control unit sends a command to the drive circuit to adjust the status indicator mechanism to a position of 2.73mm. The actual position is measured in real-time by the position sensor and is found to be 2.75mm. Error calculation: e(t) = Rm (t)-R Fusion (t) = 2.75 - 2.73418 = 0.01582 mm.

[0305] (8) Adaptive learning and parameter optimization

[0306] Based on measurement errors, the system performs adaptive parameter updates:

[0307] Learning rate settings:

[0308] η α =0.02 (learning rate adjusted according to algorithm);

[0309] η w =0.05 (Neural network weight learning rate);

[0310] η M =0.03 (learning rate of the mapping matrix).

[0311] Parameter update:

[0312] α(t+1)=α(t)+η α ·e(t)·(L-L0)=0.25+0.02·0.01582·(-0.15)=0.25-0.0000474=0.24995.

[0313] w1(t+1)=w1(t)+η w ·e(t)· =0.15+0.05·0.01582·0 (All ReLU outputs are 0)=0.15.

[0314] M 11 (t+1)=M 11 (t)+η M ·e(t)·V1=0.015+0.03·0.01582·0.82=0.015+0.00039=0.01539.

[0315] (9) Anomaly detection and safety assessment

[0316] The system performs anomaly detection and security assessment:

[0317] ① Algorithm output consistency check:

[0318] max|R i -R j | / σ R =|2.73418-1.78| / 0.5=0.95418 / 0.5=1.91>1.5 (threshold).

[0319] If an inconsistency anomaly is detected, the conflict resolution mechanism is initiated: analyze the historical accuracy and current environmental adaptability of each algorithm; increase the sampling frequency of environmental sensors; initiate a redundancy verification process; and temporarily increase the conservatism of coupling terms.

[0320] Conclusion: Under extreme low temperature conditions, the system operates normally but is approaching the warning threshold; preventative maintenance checks are recommended.

[0321] 3. Comparison of algorithm fusion effects under various working conditions

[0322] like Figure 8 , Figure 9 As shown in the table below, the algorithm weight allocation and system performance under different operating conditions are as follows:

[0323]

[0324] The data in the table clearly shows the system's adaptive capability: under standard conditions, the mechanical algorithm dominates; under high-temperature conditions, the weight of the thermal compensation algorithm increases significantly; under strong vibration conditions, the vibration suppression algorithm becomes dominant; and under combined extreme conditions, the weight of the coupling compensation term reaches its maximum, fully demonstrating the synergistic fusion effect between algorithms.

[0325] 4. Analysis of the performance improvement effect of collaborative integration on the system

[0326] Taking extreme low-temperature environments as an example, the performance comparison of three algorithm schemes is analyzed:

[0327] Using only mechanical algorithms:

[0328] R mech =ΔP mech -α·(L-L0)=1.51426-0.25·(-0.15)=1.55176mm

[0329] Actual position: 2.75mm;

[0330] Error: 2.75 - 1.55176 = 1.19824 mm (error rate 77.2%);

[0331] Three simple superposition algorithms:

[0332] R_sum=ΔP mech -α·(L-L0)-ΔP T -ΔP V =1.51426-(-0.0375)-0.0001-0.09151

[0333] =1.55176-0.0001-0.09151=1.46015mm;

[0334] Actual location: 2.75m;

[0335] Error: 2.75 - 1.46015 = 1.28985 mm (error rate 88.3%).

[0336] The synergistic fusion scheme of the present invention: R Fusion =2.73418mm; Actual position: 2.75mm; Error: 2.75-2.73418=0.01582mm (error rate 0.58%).

[0337] Analysis shows that in extreme environments, simple superposition algorithms not only fail to improve accuracy but may even increase errors. This is because simple superposition ignores the interaction between algorithms, especially when these interactions become significant under extreme conditions. The collaborative fusion scheme of this invention achieves high-precision compensation through a cross-influence model and dynamic weight adjustment.

[0338] 5. Long-term system performance evaluation

[0339] This system has been running for 12 months. By comparing the initial running data with the current data, the system's self-learning and adaptive capabilities are evaluated.

[0340]

[0341] Data shows that through continuous learning and adaptive optimization, the system not only improves accuracy but also significantly enhances its environmental adaptability while reducing computational burden. This "the more you use it, the more accurate it becomes" characteristic is a unique advantage of the collaborative fusion technology of this invention.

[0342] 6. Summary

[0343] Based on the detailed calculation process and multi-condition analysis under the above extreme low temperature environment, the following key performance evaluation conclusions can be drawn:

[0344] The accuracy is improved to ±0.02mm under standard conditions, which is 90% higher than that of traditional technology; it can still maintain an accuracy of ±0.08mm under extreme conditions, far exceeding the performance of traditional technology under extreme conditions; the calculation process verifies that the system can still maintain a low error rate of 0.58% under the low temperature of -35℃ and vibration environment.

[0345] The system can accurately identify and compensate for the effects of temperature, vibration and their interaction; the cross-influence model accurately captures the interaction between physical factors under extreme conditions; the weight of the coupling compensation term is automatically increased to 0.42 in extreme environments, reflecting the system's adaptability.

[0346] 12 months of operation data show that the system accuracy has improved by 60-68%; the environmental response time has decreased by 84-92%, significantly improving the system's reaction speed; and the parameter optimization frequency has decreased by 83%, indicating that the system is gradually converging to the optimal state.

[0347] The lead time for fault prediction has increased from 36 hours to 72 hours, providing maintenance personnel with ample response time; the anomaly detection mechanism can identify potential problems in a timely manner under extreme conditions; and the system can automatically adjust maintenance recommendations based on real-time status assessment.

[0348] Through the adaptive multi-algorithm collaborative fusion technology demonstrated in detail in this calculation process, the miniature circuit breaker switch status indication system achieves unprecedented high precision, all-condition adaptability and intelligence level while maintaining a simple structure, providing new application prospects for the intelligent development of power equipment.

Claims

1. An adaptive miniature circuit breaker switch status indication system, characterized by, The utility model relates to a kind of mechanical state monitoring and compensation system, including: mechanical structure unit, the mechanical structure unit includes vertical bar, contact point and state indicating mechanism, the vertical bar with the contact point interacts and drives the state indicating mechanism to show switch state;Sensor network unit, the sensor network unit includes position sensor, temperature sensor network and vibration sensor, wherein the temperature sensor network includes multiple distributed temperature sensors, the position sensor is used to measure switch position and state indicating position, the temperature sensor network is used to measure ambient temperature and component temperature distribution, the vibration sensor is used to monitor system vibration state;Data acquisition processing unit, the data acquisition processing unit is connected with the sensor network unit, for collecting, synchronizing and pre-processing sensor data;Algorithm processing unit, the algorithm processing unit is connected with the data acquisition processing unit, including according to algorithm module, thermal compensation algorithm module and vibration suppression algorithm module, for executing mechanical deviation correction, temperature compensation and vibration suppression calculation;Fusion decision unit, the fusion decision unit is connected with the algorithm processing unit, for integrating multiple algorithm output results, calculate final compensation value, and dynamically adjust algorithm weight according to environmental condition;Execution control unit, the execution control unit is connected with the fusion decision unit, for adjusting the position of state indicating mechanism according to the output of fusion decision unit. The mechanical deviation correction algorithm executed by the according algorithm module is: R=ΔP-α·(L-L0);Wherein, R indicates the corrected switch indicating position, unit is millimeter (mm);Alpha indicates adjustment coefficient, dimensionless;L0Indicate bar length in ideal state, unit is millimeter (mm). The fusion decision unit further includes mechanical-thermal-vibration cross-influence model, and the mechanical-thermal-vibration cross-influence model is:

2. The adaptive miniature circuit breaker switch status indication system of claim 1, wherein: The algorithm module comprises a switch position deviation algorithm model, which is: ΔP=f(L, θ, F c ); wherein ΔP represents the deviation of the switch position, in millimeters (mm); L represents the length of the vertical rod, in millimeters (mm); θ represents the angle deviation when the vertical rod contacts the contact plane, in degrees (°); and F c represents the force of the contact point, in Newton (N).

3. The adaptive miniature circuit breaker switch status indication system of claim 1, wherein: The algorithm module comprises a rod motion model in a closed state, and the rod motion model in the closed state is F c =k·(x-x0); wherein, F c represents the force of the contact point, and the unit is Newton (N); k represents the elastic coefficient of the rod, and the unit is Newton / millimeter (N / mm); x represents the displacement of the current rod, and the unit is millimeter (mm); x0 represents the initial contact position, and the unit is millimeter (mm).

4. The adaptive miniature circuit breaker switch status indication system of claim 1, wherein: ​ 5. The adaptive miniature circuit breaker switch status indication system of claim 1, wherein: The thermal compensation algorithm module includes a temperature-bias nonlinear mapping model, which is: ΔP T (t) =∑[w i ·σ(v i ·T(t) + b i )] + γ · dT(t) / dt; wherein, ΔP T (t) represents a temperature-induced time-varying bias amount, in millimeters (mm); T(t) represents a current temperature field distribution, in degrees Celsius (°C); w i ,v i ,b i represent neural network weight and bias parameters; σ represents an activation function; γ represents a temperature change rate coefficient, in millimeters / (degree Celsius·second); and dT(t) / dt represents a temperature change rate, in degrees Celsius / second (°C / s).

6. The adaptive miniature circuit breaker switch status indication system of claim 1, wherein: The thermal compensation algorithm module comprises a material thermal expansion nonlinear coefficient model, which is: a T (T) = a0 + a1 · T + a2 · T 2 + a3 · exp(T / T c ); wherein, a T (T) represents a temperature-dependent thermal expansion coefficient, with a unit of 1 / degree Celsius (1 / ℃); a0, a1, a2, a3 represent fitting coefficients; T represents a current temperature, with a unit of degrees Celsius (℃); T c represents a material critical temperature coefficient, with a unit of degrees Celsius (℃).

7. The adaptive miniature circuit breaker switch status indication system of claim 1, wherein: The vibration suppression algorithm module includes a vibration-bias mapping matrix model, which is: ΔP V = M · V + V^T · N · V; where ΔP V represents a vibration-induced bias vector, with units of millimeters (mm); M represents a linear mapping matrix; N represents a second-order nonlinear mapping tensor; V represents a vibration feature vector, containing vibration frequency and amplitude information; V^T represents the transpose of V.

8. The adaptive miniature circuit breaker switch status indication system of claim 1, wherein: The fusion decision unit comprises a unified state indication deviation compensation model, which is ΔP total (t)=ΔP mech (t)+ΔP therm (t)+ΔP vib (t)+ΔP coupled (t);wherein, ΔP total (t) represents a total deviation amount, in millimeters (mm); ΔP mech (t) represents a deviation according to a mechanical assembly; ΔP therm (t) represents a deviation caused by temperature change; ΔP vib (t) represents a deviation caused by vibration; ΔP coupled (t) represents an interaction deviation caused by multi-factor coupling.

9. The adaptive miniature circuit breaker switch status indication system of claim 8, wherein: ​ ΔP coupled (t) = λ TM · f TM (T, L, k) + λ VM · f VM (V, F c , x) + λ TV · f TV (T, V); wherein λ TM , λ VM , λ TV represent coupling coefficients; f TM (T, L, k) represents a temperature influence function on mechanical properties; f VM (V, F c , x) represents a vibration influence function on contact force and displacement; and f TV (T, V) represents a temperature and vibration interaction function.

Images