Test system, method and equipment based on artificial intelligence
Through the artificial intelligence-based test system, using fault tree analysis and physical model construction, the cause of motor failure is gradually decomposed, and the problems of single motor life testing methods and sensor impact in the existing technology are solved, achieving more accurate life prediction and detection effects .
Patent Information
- Application Number
- CN202411173160.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-26
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing motor life test method is single, and it is difficult to effectively test in complex environments. The sensor installation will affect the normal operation of the motor, resulting in inconsistent test results with the actual situation.
Adopt an artificial intelligence-based test system, including parameter determination module, parameter classification module, status labeling module and life prediction module, through fault collection, fault tree analysis, physical model construction and logistic regression model training, gradually decompose the cause of the fault and improve parameter detection Accuracy, reduce dependence on sensors, and achieve life expectancy.
In-depth analysis of motor failure mode is realized, parameter detection accuracy is improved, and interference to motor operation is reduced. It is especially suitable for motor detection in less contact or high-risk environments, providing more accurate life prediction.
Smart Images

Figure CN119024162A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a testing system, method and equipment based on artificial intelligence. Background Art
[0002] Motor life testing is an important part of ensuring motor quality and reliability. Life testing can identify key factors that affect motor life, such as material selection, structural design, and manufacturing process, by analyzing the motor's performance and life under different working conditions. Based on these feedbacks, engineers can optimize the motor's design, select more suitable materials and components, improve the manufacturing process, and improve the motor's overall performance and life, thus pushing motor technology to a higher level.
[0003] Common motor life test methods include: operating time test, accelerated life test, temperature cycle test, vibration monitoring and insulation resistance test; however, the variables of the above test methods are relatively single, and it is difficult to effectively test the motor life in a more complex environment; in addition, most testing technologies require the use of sensors, but the installation of sensors will affect the normal operation of the motor system, resulting in inconsistencies with actual operating conditions. Summary of the invention
[0004] The purpose of the present invention is to provide a testing system, method and device based on artificial intelligence to solve the problems raised in the prior art.
[0005] To achieve the above object, the present invention provides the following technical solutions: a test system based on artificial intelligence, the system comprising a parameter determination module, a parameter classification module, a state marking module and a life prediction module;
[0006] The parameter determination module is used to determine the parameters required for motor life prediction; the parameter classification module is used to classify the collected parameters and calculate the parameter values according to the feasibility of sensor collection; the state standard module is used to mark the state of the records corresponding to the parameters and their values; the life prediction module is used to classify the new parameter state and predict the motor life according to the parameter and its corresponding state training model;
[0007] The output end of the parameter determination module is connected to the input end of the parameter classification module; the output end of the parameter classification module is connected to the input end of the state marking module; the output end of the state marking module is connected to the input end of the life prediction module.
[0008] The parameter determination module includes a fault collection unit, a fault tree unit and a parameter determination unit;
[0009] The fault collection unit is used to collect the fault mode of the motor; the fault tree unit is used to trace the cause of the fault according to the fault tree analysis method; the parameter determination unit is used to determine the parameters to be collected according to the tracing result;
[0010] The output end of the fault collection unit is connected to the input end of the fault tree unit; the output end of the fault tree unit is connected to the input end of the parameter determination unit; the output end of the parameter determination unit is connected to the input end of the parameter classification module.
[0011] The parameter classification module includes a parameter classification unit, a sensor unit, a direct parameter unit, an indirect parameter unit, a physical model unit and a parameter merging unit;
[0012] The parameter classification unit is used to classify the determined parameters according to the feasibility of sensor measurement; the sensor unit is used to collect different types of data; the direct parameter unit is used to store parameters directly obtained by the sensor; the indirect parameter unit is used to store parameters that cannot be directly obtained by the sensor; the physical model unit is used to construct a physical model to calculate the indirect parameters; the parameter merging unit is used to merge the direct parameters and the indirect parameters obtained by calculation;
[0013] The output end of the parameter classification unit is connected to the input end of the sensor unit; the output end of the sensor unit is connected to the input ends of the direct parameter unit and the indirect parameter unit; the output end of the direct parameter unit is connected to the input end of the parameter merging unit; the output end of the indirect parameter unit is connected to the input end of the physical model unit; the output end of the physical model unit is connected to the input end of the parameter merging unit; the output end of the parameter merging unit is connected to the input end of the state labeling module.
[0014] The state marking module includes a fault recording unit, a state classification unit and a state marking unit;
[0015] The fault recording unit is used to provide a basis for the classification and labeling of the state; the state classification unit is used to classify the state; the state labeling unit is used to label the state of the parameter;
[0016] The output end of the fault recording unit is connected to the input end of the state classification unit; the output end of the state classification unit is connected to the input end of the state marking unit; the output end of the state marking unit is connected to the input end of the life prediction module.
[0017] The life prediction module includes a data processing unit, a model building unit, a model classification unit and a life prediction unit;
[0018] The data processing unit is used to standardize parameter data and label state classification; the model building unit is used to build and train a model with parameters as features and states as targets; the model classification unit is used to predict the state of new parameters according to the model; the life prediction unit is used to determine the life of the motor according to the predicted state;
[0019] The output end of the data processing unit is connected to the input end of the model building unit; the output end of the model building unit is connected to the input end of the model classification unit; the output end of the model classification unit is connected to the input end of the life prediction unit.
[0020] The testing method based on artificial intelligence comprises the following steps:
[0021] Step 1: Collect the motor failure modes, identify the potential causes of motor failure, and determine the detection parameters;
[0022] Step 2: Divide the parameters into direct parameters and indirect parameters. For indirect parameters, determine the indirect detection method.
[0023] Step 3: The operator marks the corresponding status of the parameter data based on the existing fault records and experience;
[0024] Step 4: Process the parameter data, establish a life prediction model based on the parameters and their corresponding states, and predict the new parameter data.
[0025] In step 1, the failure mode is obtained based on historical failure data, failure reports and literature research, and is expressed as: [F1, F2, …, F m ];
[0026] Wherein, m is a positive integer, indicating the type of failure mode; F m represents the mth failure mode;
[0027] The specific method of identifying the potential causes of motor failure is the fault tree analysis method:
[0028] S7-1. Define each failure mode as a top event;
[0029] S7-2, analyzing the direct cause of the top event: decomposing the direct cause of the top event into multiple sub-events;
[0030] S7-3, further decomposing sub-events: continue to decompose sub-events until they cannot be further decomposed or specific parameters have been reached;
[0031] S7-4, determine the detection parameters: Based on the in-depth analysis of the failure mode, identify the detection parameters of each potential failure cause, expressed as [P1, P2, ..., P n ];
[0032] Where n is a positive integer, indicating the type of detection parameter; P n Indicates the nth detection parameter.
[0033] In step 2, the parameters are classified into direct parameters and indirect parameters based on the feasibility of sensor measurement; the direct parameters are directly obtained by the sensor and are expressed as [D1, D2, ..., D i ]; the indirect parameters cannot be directly obtained by the sensor, but are obtained by indirect detection, which is expressed as [I1, I2, …, I j ];
[0034] Where i is a positive integer, indicating the type of direct parameter; D i represents the i-th direct parameter; j is a positive integer, indicating the type of indirect parameter; I j Indicates the jth indirect parameter; i+j=n;
[0035] The specific method of the indirect detection is:
[0036] S8-1. Establish physical model: Establish a model that reflects the dynamic behavior of the motor; the model consists of differential equations, algebraic equations or difference equations, describing the relationship between input, output and state variables;
[0037] S8-2. Define input parameters and output parameters: Input parameters are parameters directly measured by sensors, and output parameters are indirect parameters, which cannot be directly measured but need to be indirectly inferred;
[0038] S8-3, Model equation establishment: Use physical laws to establish differential equations or algebraic equations that describe the relationship between input parameters and output parameters;
[0039] S8-4, solving the model equation: obtaining the indirect parameter value by solving the model equation;
[0040] In step 3, the operator classifies and labels the corresponding states of the parameter data according to the existing fault records and experience; the state classification is expressed as: [K1, K2, …, K p ]; The motor life is defined as p levels, then for the corresponding level of K1: the life is 1; for the corresponding level of K2: the life is (p-2) / (p-1); ...; for K p-1 Corresponding level: Lifespan is (p-(p-1)) / (p-1); for K p Corresponding level: life span is (pp) / (p-1)=0;
[0041] Where p is a positive integer, p>1, indicating the type of state; K p represents the p-th type of state;
[0042] A complete parameter and status record is represented as: [D1, D2, ..., D i ,I1,I2,…,I j ,K u ];
[0043] Among them, u is a positive integer, K u ∈{K1,K2,…,K p};
[0044] In step 4, the parameter data is processed as follows: the parameter data is standardized to have the same scale; for the parameter data of the same motor at different times, the data is segmented using a sliding window to capture the time correlation; different states in the record are represented by numbers: [1: K1, 2: K2, …, p: K p ];
[0045] The specific method of establishing the life prediction model according to the parameters and their corresponding states is: using a logistic regression model to predict the state;
[0046] Model input: X = [V(D1), V(D2), …, V(D i ),V(I1),V(I2),…,V(I j )];
[0047] Model output: probability distribution of state Y, denoted as G(Y=k|X), which represents the probability that the motor belongs to state category k given the input parameter X;
[0048] Among them, V(D i ) represents the i-th direct parameter D i The standardized value; V(I j ) represents the jth indirect parameter I j The standardized value; k represents the category number of the state, k∈{1,2,…,p};
[0049] There are p states, and the softmax regression model is defined as:
[0050]
[0051] Among them, θ k is a parameter vector representing the logistic regression model parameters for category k;
[0052] The cross entropy loss function L(θ) is used to measure the difference between the model prediction and the true label:
[0053]
[0054] Where N is the total number of samples, y s,k is the actual label of sample s corresponding to category k, y s,k 1 means the sth sample belongs to category k, otherwise it is 0; X s Represents the input parameter of the sth sample;
[0055] The loss function is optimized by the gradient descent algorithm to minimize the loss function; after training, the state category of the new input X is predicted by maximizing the probability G(Y=k|X)
[0056]
[0057] The motor life is determined based on the level corresponding to the predicted category.
[0058] The test equipment based on artificial intelligence includes the test system and the test method.
[0059] Compared with the prior art, the present invention has the following beneficial effects:
[0060] 1. The present invention gradually decomposes the fault cause through hierarchical analysis of the fault mode. The hierarchical analysis ensures the depth and breadth of the fault analysis and can more accurately identify the potential fault cause;
[0061] 2. The present invention improves the accuracy of parameter detection by establishing a physical model that reflects the dynamic behavior of the motor and inferring the indirect parameter value using differential equations or algebraic equations, thereby reducing the reliance on expensive or difficult to install direct measurement equipment;
[0062] 3. The present invention does not rely on direct contact with the motor for detection and life prediction, which reduces interference with the normal operation of the motor and is particularly suitable for motor detection in difficult-to-access or high-risk environments. BRIEF DESCRIPTION OF THE DRAWINGS
[0063] Figure 1 It is a flow chart of the test system based on artificial intelligence of the present invention;
[0064] Figure 2 It is a schematic diagram of the steps of the artificial intelligence-based testing method of the present invention. DETAILED DESCRIPTION
[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0066] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a testing method based on artificial intelligence, which includes the following steps:
[0067] Step 1: Collect the motor failure modes, identify the potential causes of motor failure, and determine the detection parameters;
[0068] The failure modes are obtained from historical failure data, failure reports and literature research, and are expressed as: [winding short circuit failure, rotor imbalance];
[0069] The specific method of identifying the potential causes of motor failure is the fault tree analysis method:
[0070] S7-1. Define each failure mode as a top event;
[0071] The top event is defined as a winding short-circuit fault; the direct cause of the top event is analyzed, and the direct cause of the winding short-circuit fault is decomposed into multiple sub-events: sub-event 1: insulation aging, sub-event 2: overcurrent, sub-event 3: overtemperature;
[0072] S7-2, analyzing the direct cause of the top event: decomposing the direct cause of the top event into multiple sub-events;
[0073] S7-3, further decomposing sub-events: continue to decompose sub-events until they cannot be further decomposed or specific parameters have been reached;
[0074] S7-4. Determine detection parameters: Based on the in-depth analysis of the failure mode, identify the detection parameters for each potential failure cause;
[0075] The top event is defined as a winding short-circuit fault;
[0076] Analyze the direct cause of the top event and decompose the direct cause of the winding short-circuit fault into multiple sub-events: sub-event 1: insulation aging, sub-event 2: overcurrent, sub-event 3: overtemperature;
[0077] Further decomposition of sub-events: Insulation aging sub-event 1: long-term operation leads to insulation material degradation, sub-event 2: frequent motor start and stop; overcurrent sub-event 1: sudden increase in load, sub-event 2: abnormal power supply voltage; S3: overtemperature sub-event 1: cooling system failure, sub-event 2: excessive ambient temperature;
[0078] Determine the detection parameters: winding temperature, vibration signal;
[0079] The top event is defined as rotor unbalance;
[0080] Analyze the direct cause of the top event and decompose the direct cause of rotor imbalance into multiple sub-events: Sub-event 1: rotor defect, Sub-event 2: improper installation, Sub-event 3: rotor damage
[0081] Further decomposition of sub-events: rotor defect sub-event 1: uneven material distribution during rotor manufacturing, sub-event 2: rotor surface damage; improper installation sub-event 1: rotor not properly aligned, sub-event 2: bearings not properly installed during installation; rotor damage sub-event 1: rotor wear caused by long-term operation, sub-event 2: rotor deformation caused by external impact;
[0082] Determine the detection parameters: vibration signal, sound signal;
[0083] The detection parameters are comprehensively determined as [winding temperature, vibration signal, sound signal];
[0084] Step 2: Divide the parameters into direct parameters and indirect parameters. For indirect parameters, determine the indirect detection method.
[0085] The classification of the parameters into direct parameters and indirect parameters is based on the feasibility of sensor measurement; the direct parameters are directly obtained by the sensor and are expressed as [vibration signal, sound signal]; the indirect parameters cannot be directly obtained by the sensor and are obtained by indirect detection, and are expressed as [winding temperature];
[0086] The specific method of the indirect detection is: constructing a physical model of the winding temperature according to the heat balance equation:
[0087]
[0088] T w is the winding temperature; In is the sensor measurement current; T ambient is the ambient temperature; winding resistance R winding ; Heat dissipation coefficient h; Heat dissipation surface area A;
[0089] Solve the model equations to obtain indirect parameter values;
[0090] Step 3: The operator marks the corresponding status of the parameter data based on the existing fault records and experience;
[0091] The operator classifies and marks the corresponding states of the parameter data according to the existing fault records and experience; the state classification is expressed as: [good, abnormal, emergency stop]; the motor life is defined as 3 levels, then for the good corresponding level: life is 1; for the abnormal corresponding level: life is 1 / 2; for the emergency stop corresponding level: life is 0;
[0092] A complete parameter and status record is represented as: [winding temperature, vibration signal, sound signal, intact / abnormal / emergency shutdown];
[0093] Step 4: Process the parameter data, establish a life prediction model based on the parameters and their corresponding states, and predict the new parameter data;
[0094] The method of processing the parameter data is as follows: standardizing the parameter data to make them have the same scale; for the parameter data of the same motor at different times, segmenting the data using a sliding window to capture the time correlation; using numbers to represent different states in the record: [1: intact, 2: abnormal, 3: emergency stop];
[0095] The specific method of establishing the life prediction model according to the parameters and their corresponding states is as follows: using a logistic regression model to realize state prediction; using a cross entropy loss function to measure the difference between the model prediction and the true label; optimizing the loss function through a gradient descent algorithm to minimize the loss function; after training, predicting the state category of the new input by maximizing the probability; and judging the motor life according to the level corresponding to the predicted category.
[0096] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.
Claims
1. The test system based on artificial intelligence is characterized by: The system includes a parameter determination module, a parameter classification module, a state marking module and a life prediction module; The parameter determination module is used to determine the parameters required for motor life prediction; the parameter classification module is used to classify the collected parameters and calculate the parameter values according to the feasibility of sensor collection; the state standard module is used to mark the state of the records corresponding to the parameters and their values; the life prediction module is used to classify the new parameter state and predict the motor life according to the parameter and its corresponding state training model; The output end of the parameter determination module is connected to the input end of the parameter classification module; the output end of the parameter classification module is connected to the input end of the state marking module; the output end of the state marking module is connected to the input end of the life prediction module.
2. The artificial intelligence-based testing system according to claim 1, characterized in that: The parameter determination module includes a fault collection unit, a fault tree unit and a parameter determination unit; The fault collection unit is used to collect the fault mode of the motor; the fault tree unit is used to trace the cause of the fault according to the fault tree analysis method; The parameter determination unit is used to determine the parameters to be collected according to the traceability results; The output end of the fault collection unit is connected to the input end of the fault tree unit; the output end of the fault tree unit is connected to the input end of the parameter determination unit; the output end of the parameter determination unit is connected to the input end of the parameter classification module.
3. The artificial intelligence-based testing system according to claim 2, characterized in that: The parameter classification module includes a parameter classification unit, a sensor unit, a direct parameter unit, an indirect parameter unit, a physical model unit and a parameter merging unit; The parameter classification unit is used to classify the determined parameters according to the feasibility of sensor measurement; the sensor unit is used to collect different types of data; The direct parameter unit is used to store parameters directly obtained by the sensor; the indirect parameter unit is used to store parameters that cannot be directly obtained by the sensor; the physical model unit is used to construct a physical model to calculate the indirect parameters; the parameter merging unit is used to merge the direct parameters and the indirect parameters obtained by calculation; The output end of the parameter classification unit is connected to the input end of the sensor unit; the output end of the sensor unit is connected to the input ends of the direct parameter unit and the indirect parameter unit; the output end of the direct parameter unit is connected to the input end of the parameter merging unit; The output end of the indirect parameter unit is connected to the input end of the physical model unit; the output end of the physical model unit is connected to the input end of the parameter merging unit; the output end of the parameter merging unit is connected to the input end of the state marking module.
4. The artificial intelligence-based testing system according to claim 3, characterized in that: The state marking module includes a fault recording unit, a state classification unit and a state marking unit; The fault recording unit is used to provide a basis for the classification and labeling of the state; the state classification unit is used to classify the state; the state labeling unit is used to label the state of the parameter; The output end of the fault recording unit is connected to the input end of the state classification unit; the output end of the state classification unit is connected to the input end of the state marking unit; the output end of the state marking unit is connected to the input end of the life prediction module.
5. The artificial intelligence-based testing system according to claim 4, characterized in that: The life prediction module includes a data processing unit, a model building unit, a model classification unit and a life prediction unit; The data processing unit is used to standardize parameter data and label state classification; the model building unit is used to build and train a model with parameters as features and states as targets; the model classification unit is used to predict the state of new parameters according to the model; the life prediction unit is used to determine the life of the motor according to the predicted state; The output end of the data processing unit is connected to the input end of the model building unit; the output end of the model building unit is connected to the input end of the model classification unit; the output end of the model classification unit is connected to the input end of the life prediction unit.
6. The test method based on artificial intelligence is characterized by: The method comprises the following steps: Step 1: Collect the motor failure modes, identify the potential causes of motor failure, and determine the detection parameters; Step 2: Divide the parameters into direct parameters and indirect parameters. For indirect parameters, determine the indirect detection method. Step 3: The operator marks the corresponding status of the parameter data based on the existing fault records and experience; Step 4: Process the parameter data, establish a life prediction model based on the parameters and their corresponding states, and predict the new parameter data.
7. The artificial intelligence-based testing method according to claim 6, characterized in that: In step 1, the failure mode is obtained based on historical failure data, failure reports and literature research, and is expressed as: [F1, F2, …, F m ]; Wherein, m is a positive integer, indicating the type of failure mode; F m represents the mth failure mode; The specific method of identifying the potential causes of motor failure is the fault tree analysis method: S7-1. Define each failure mode as a top event; S7-2, analyzing the direct cause of the top event: decomposing the direct cause of the top event into multiple sub-events; S7-3, further decomposing sub-events: continue to decompose sub-events until they cannot be further decomposed or specific parameters have been reached; S7-4, determine the detection parameters: Based on the in-depth analysis of the failure mode, identify the detection parameters of each potential failure cause, expressed as [P1, P2, ..., P n ]; Where n is a positive integer, indicating the type of detection parameter; P n Indicates the nth detection parameter.
8. The artificial intelligence-based testing method according to claim 7, characterized in that: In step 2, the parameters are classified into direct parameters and indirect parameters based on the feasibility of sensor measurement; the direct parameters are directly obtained by the sensor and are expressed as [D1, D2, ..., D i ]; the indirect parameters cannot be directly obtained by the sensor, but are obtained by indirect detection, which is expressed as [I1, I2, …, I j ]; Where i is a positive integer, indicating the type of direct parameter; D i represents the i-th direct parameter; j is a positive integer, indicating the type of indirect parameter; I j Indicates the jth indirect parameter; i+j=n; The specific method of the indirect detection is: S8-1. Establish physical model: Establish a model that reflects the dynamic behavior of the motor; the model consists of differential equations, algebraic equations or difference equations, describing the relationship between input, output and state variables; S8-2. Define input parameters and output parameters: Input parameters are parameters directly measured by sensors, and output parameters are indirect parameters, which cannot be directly measured but need to be indirectly inferred; S8-3, Model equation establishment: Use physical laws to establish differential equations or algebraic equations that describe the relationship between input parameters and output parameters; S8-4. Solve the model equation: obtain the indirect parameter value by solving the model equation.
9. The artificial intelligence-based testing method according to claim 8, characterized in that: In step 3, the operator classifies and labels the corresponding states of the parameter data according to the existing fault records and experience; the state classification is expressed as: [K1, K2, …, K p ]; The motor life is defined as p levels, then for the corresponding level of K1: the life is 1; for the corresponding level of K2: the life is (p-2) / (p-1); ...; for K p-1 Corresponding level: Lifespan is (p-(p-1)) / (p-1); for K p Corresponding level: life span is (pp) / (p-1)=0; Where p is a positive integer, p>1, indicating the type of state; K p represents the p-th type of state; A complete parameter and status record is represented as: [D1, D2, ..., D i ,I1,I2,…,I j ,K u ]; Among them, u is a positive integer, K u ∈{K1,K2,…,K p }; In step 4, the parameter data is processed as follows: the parameter data is standardized to have the same scale; for the parameter data of the same motor at different times, the data is segmented using a sliding window to capture the time correlation; different states in the record are represented by numbers: [1: K1, 2: K2, …, p: K p ]; The specific method of establishing the life prediction model according to the parameters and their corresponding states is: using a logistic regression model to predict the state; Model input: X = [V(D1), V(D2), …, V(D i ),V(I1),V(I2),…,V(I j )]; Model output: probability distribution of state Y, denoted as G(Y=k|X), which represents the probability that the motor belongs to state category k given the input parameter X; Among them, V(D i ) represents the i-th direct parameter D i The standardized value; V(I j ) represents the jth indirect parameter I j The standardized value; k represents the category number of the state, k∈{1,2,…,p}; There are p states, and the softmax regression model is defined as: Among them, θ k is a parameter vector representing the logistic regression model parameters for category k; The cross entropy loss function L(θ) is used to measure the difference between the model prediction and the true label: Where N is the total number of samples, y s,k is the actual label of sample s corresponding to category k, y s,k 1 means the sth sample belongs to category k, otherwise it is 0; X s Represents the input parameter of the sth sample; The loss function is optimized by the gradient descent algorithm to minimize the loss function; after training, the state category of the new input X is predicted by maximizing the probability G(Y=k|X) The motor life is determined based on the level corresponding to the predicted category.
10. Test equipment based on artificial intelligence, characterized by: It comprises the test system according to any one of claims 1 to 5 and the test method according to any one of claims 6 to 9.
