A Data-Based Method and System for Fault Diagnosis of Air Conditioners in Mining Trucks

Abstract

This application relates to the field of air conditioning fault diagnosis technology, and discloses a data analysis-based method and system for diagnosing faults in mining truck air conditioners. The method includes: acquiring environmental data from the mining truck and the current detection values ​​of key parameters in sensors; determining whether the performance of the mining truck air conditioner has degraded; if so, determining whether the key parameters are abnormal; if abnormalities occur, recording the time from the most recent start of normal operation of the mining truck air conditioner to the occurrence of the abnormality, and the key parameters involved; acquiring the causes of faults in the mining truck air conditioner within each historical reference period; calculating the first probability of the target parameter becoming abnormal under each fault cause condition, and the second probability caused by each fault cause when a component of the mining truck air conditioner fails; calculating the third probability of each fault cause based on the first and second probabilities, and determining the fault maintenance sequence and optimal maintenance time. This application improves the accuracy of fault diagnosis for mining truck air conditioners.

Description

Technical Field

[0001] This application relates to the field of air conditioning fault diagnosis technology, and in particular to a method and system for fault diagnosis of mining truck air conditioning based on data analysis. Background Technology

[0002] Mining trucks are key equipment in mine transportation, and their operational efficiency and reliability directly impact the safety and economic benefits of mine production. The air conditioning system of these trucks, as a crucial component ensuring driver comfort and normal equipment operation, has always been a key focus and challenge in mine equipment management due to its fault diagnosis and maintenance. Because of the harsh mining environment, such as high temperatures, dust, and vibration, mining truck air conditioning systems face a higher risk of failure. A malfunction in the air conditioning system can not only affect the driver's operational state but also lead to equipment overheating, impacting the overall performance of the mining truck and even causing safety accidents.

[0003] Existing patent applications, such as Chinese patent application CN104635719A, disclose a method, device, and system for diagnosing automotive air conditioning faults. This includes: detecting the pressure switch signal of a three-state pressure switch in the air conditioning system and the ambient temperature; and diagnosing whether a fault has occurred in the air conditioning system based on the pressure switch signal and the ambient temperature.

[0004] For example, Chinese patent application CN104635719A discloses a method and related components for diagnosing faults in train air conditioning units. This method involves determining the first time required for the air conditioning unit's air supply temperature to change to a preset temperature value. Using this first time as the basis for fault diagnosis allows for prediction of potential malfunctions before the unit generates fault data, thus avoiding adverse effects on the unit's lifespan. A fault is determined when the first time exceeds a final alarm time. The final alarm time is an adjusted value based on whether the air conditioning unit reports fault data, further ensuring the accuracy of fault prediction and simplifying the detection method.

[0005] However, both of the aforementioned existing technologies lack multi-source data fusion and predictive analysis, and cannot pinpoint the cause of the fault. Therefore, this application proposes a data analysis-based fault diagnosis method and system for mining truck air conditioners. Summary of the Invention

[0006] To address the aforementioned technical issues, this application provides a data analysis-based fault diagnosis method and system for mining truck air conditioners, which aims to improve the accuracy of fault diagnosis for mining truck air conditioners.

[0007] Firstly, this application provides a data analysis-based method for diagnosing faults in mining truck air conditioners, the method comprising:

[0008] Step S1: Deploy multiple sensors on the relevant components of the mining truck air conditioner to obtain environmental data of the mining truck and the current detection values ​​of key parameters in the sensors. Determine whether the performance of the mining truck air conditioner has degraded at preset intervals. If so, determine whether the key parameters have become abnormal.

[0009] Step S2: If an anomaly occurs, record the time from the most recent start of normal operation of the mining truck air conditioner to the occurrence of the anomaly, as well as the key parameters of the anomaly, which are defined as the first time and the target parameter, respectively. Based on the first time, extract multiple historical reference time periods from the preset historical fault data table.

[0010] Step S3: Obtain the causes of failure of the mining truck air conditioner in each historical reference period, calculate the first probability of the target parameter becoming abnormal under each failure cause, and the second probability caused by each failure cause when the components of the mining truck air conditioner fail.

[0011] Step S4: Calculate the third probability of each fault cause when the target parameter is abnormal based on the first probability and the second probability, determine the fault maintenance sequence of the mining air conditioner based on the third probability, and calculate the optimal maintenance time.

[0012] In conjunction with the first aspect, in the first implementation of the first aspect of this application, determining whether the performance of the mining truck air conditioner has degraded at preset time intervals includes:

[0013] The environmental data includes the initial temperature and current temperature inside the cab of the mining truck. The mining truck air conditioner has multiple operating modes, each corresponding to an initial temperature range, a target temperature, and a baseline adjustment time. The baseline adjustment time is the average time required to adjust from the historical initial temperature inside the cab to the corresponding target temperature under normal operating conditions of the mining truck air conditioner. The current operating mode of the mining truck air conditioner is obtained based on the initial temperature.

[0014] Determine whether the current temperature has reached the target temperature corresponding to the current operating mode. If it has, record the time required to adjust from the starting temperature to the target temperature, which is defined as the current adjustment time. Calculate the first deviation degree of the mining air conditioner based on the current adjustment time and the reference adjustment time. If the first deviation degree is greater than a first threshold, determine that the mining air conditioner has deteriorated.

[0015] In conjunction with the first aspect, in the second implementation of the first aspect of this application, determining whether the key parameter is abnormal includes:

[0016] A normal prediction model is created for each sensor as the target sensor. The historical detection values ​​of key parameters of all other sensors under normal conditions are used as input data for the normal prediction model. The normal prediction model learns the correlation between multiple sensors under normal conditions and outputs the predicted value of the target sensor under normal conditions. The second deviation degree between the current detection value and the predicted value of the key parameter of each target sensor is calculated. It is determined whether there is a second deviation degree greater than a second threshold. If there is no deviation degree, it means that the sensor of the mining air conditioner is not abnormal. If there is a deviation degree, it is determined that the key parameter corresponding to the second deviation degree greater than the second threshold is abnormal.

[0017] In conjunction with the first aspect, in the third implementation of the first aspect of this application, multiple historical reference time periods are extracted from a preset historical fault data table based on the first time, including:

[0018] The historical fault data table includes the historical cycle of the mining air conditioner from the end of the previous maintenance cycle to the occurrence of an anomaly and then to the end of the current maintenance cycle, the name of the component that failed, the cause of the failure, and the key parameters of the anomaly at the time of the failure. For each historical cycle, an equivalent starting point is calculated. The equivalent starting point is the time point after adding the first time to the end time of the previous maintenance cycle. The time period from the equivalent starting point to the end of the current maintenance cycle is defined as the historical reference period.

[0019] In conjunction with the first aspect, in the fourth implementation of the first aspect of this application, calculating the first probability that the target parameter will become abnormal under each fault cause condition includes:

[0020] The first number of times each fault cause caused the mining truck air conditioner to fail in the historical fault data table and the second number of times the key parameter of the abnormality corresponding to the fault cause occurred are obtained as the target parameter. The ratio of the second number to the first number is defined as the first probability.

[0021] In conjunction with the first aspect, in the fifth implementation of the first aspect of this application, the calculation of the second probability caused by each cause of failure in the case of a component failure in the mining truck air conditioner includes:

[0022] Obtain the third number of times each component fails and the fourth number of times each component fails due to the corresponding fault cause from the historical fault data table, and define the ratio of the fourth number to the third number as the second probability.

[0023] In conjunction with the first aspect, in the sixth implementation of the first aspect of this application, a third probability of each fault cause occurring when the target parameter becomes abnormal is calculated based on the first probability and the second probability, including:

[0024] The sum of the products of the first probability and the second probability corresponding to all fault causes is defined as the total probability of the target parameter becoming abnormal. The ratio of the product of the first probability and the second probability corresponding to each fault cause to the total probability is defined as the third probability of each fault cause occurring when the target parameter becomes abnormal.

[0025] In conjunction with the first aspect, in the seventh implementation of the first aspect of this application, determining the fault maintenance sequence of the mining truck air conditioner based on the third probability includes:

[0026] Based on the historical fault data table, obtain the first distribution function corresponding to each fault cause. The first distribution function is the distribution function of the historical probability of the mining air conditioner failing before the key parameters are abnormal under the action of the fault cause, as a function of the running time. Substitute the first time into the first distribution function to obtain the first function value. The product of the third probability corresponding to each fault cause and the first function value is defined as the fault probability of the mining air conditioner failing due to the fault cause within the first time period. Perform fault maintenance on the mining air conditioner based on each fault cause in descending order of the fault probability.

[0027] In conjunction with the first aspect, in the eighth implementation of the first aspect of this application, the calculation of the optimal maintenance time includes:

[0028] The sum of the failure probabilities corresponding to all failure causes is defined as the total probability of the mining truck air conditioner failing within the first time period. As the mining truck air conditioner runs, it is determined at preset intervals whether the total probability is greater than a third threshold. If it is, the running time of the mining truck air conditioner at this time is defined as the second time point, and the time period between the first time point and the second time point is defined as the optimal maintenance time.

[0029] Secondly, this application provides a data analysis-based fault diagnosis system for mining truck air conditioners, the system comprising:

[0030] The monitoring module is used to deploy multiple sensors on the relevant components of the mining truck air conditioner to acquire environmental data of the mining truck and the current detection values ​​of key parameters in the sensors. It determines whether the performance of the mining truck air conditioner has deteriorated at preset intervals, and if so, determines whether the key parameters have become abnormal.

[0031] The extraction module is used to record the time from the most recent start of normal operation of the mining truck air conditioner to the occurrence of the abnormality, as well as the key parameters of the abnormality, if an abnormality occurs. These are defined as the first time and the target parameter, respectively. Based on the first time, multiple historical reference time periods are extracted from a preset historical fault data table.

[0032] The calculation module obtains the causes of failure of the mining truck air conditioner in each historical reference period, calculates the first probability of the target parameter becoming abnormal under each failure cause, and the second probability caused by each failure cause when the components of the mining truck air conditioner fail.

[0033] The maintenance module is used to calculate the third probability of each fault cause when the target parameter is abnormal, based on the first probability and the second probability, determine the fault maintenance sequence of the mining air conditioner based on the third probability, and calculate the optimal maintenance time.

[0034] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0035] This application achieves real-time monitoring and fault warning of mining truck air conditioning performance by deploying sensors on relevant components to acquire real-time environmental data and current detection values ​​of key parameters. Based on preset time intervals, it determines whether air conditioning performance has degraded. Furthermore, by recording the time of anomalies and key parameters, it extracts reference time periods from historical fault data tables and calculates the probability of each fault causing abnormal target parameters and the probability of each component's fault being caused by each fault cause. This determines the fault maintenance sequence and optimal maintenance time, providing clear guidance for maintenance personnel and improving maintenance efficiency and accuracy. In addition, this application further improves fault diagnosis accuracy by creating a normal prediction model to learn the correlation between sensors under normal conditions. In summary, this application utilizes a multi-dimensional probability model to quantify fault probability, dynamically adjust the maintenance sequence and optimal maintenance time, significantly reducing misdiagnosis rates, improving maintenance efficiency, reducing repair costs, and extending equipment lifespan. Attached Figure Description

[0036] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram of an embodiment of the data analysis-based fault diagnosis method for mining truck air conditioners in this application.

[0038] Figure 2 This is a schematic diagram of the historical fault data table in an embodiment of this application;

[0039] Figure 3 This is a schematic diagram of historical parameter time periods in the embodiments of this application;

[0040] Figure 4This is a schematic diagram of one embodiment of the data analysis-based fault diagnosis system for mining truck air conditioners in this application. Detailed Implementation

[0041] This application provides a data analysis-based method and system for diagnosing faults in mining air conditioning systems. The terms "first," "second," "third," "fourth," etc. (if present)," in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0042] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the data analysis-based fault diagnosis method for mining air conditioners in this application includes:

[0043] Step S1: Deploy multiple sensors on the relevant components of the mining truck air conditioner to obtain environmental data of the mining truck and the current detection values ​​of key parameters in the sensors. Determine whether the performance of the mining truck air conditioner has deteriorated at preset intervals. If so, determine whether the key parameters have become abnormal.

[0044] Specifically, mining trucks typically operate in extreme environments (such as high temperature, high dust, and high vibration), requiring their air conditioning systems to continuously provide stable temperature control for the cab. However, air conditioning performance degradation (such as decreased cooling efficiency and increased energy consumption) can be caused by component aging, environmental degradation, or untimely maintenance, directly affecting driver comfort and equipment lifespan. To address these issues, this application deploys sensors on key components of the mining truck's air conditioning system to monitor environmental data and parameter changes in real time. For example, environmental sensors such as temperature and humidity sensors, dust sensors, and vibration sensors are deployed in the mining truck's cab, air conditioning inlet, and engine compartment to collect key parameters such as ambient temperature, humidity, dust concentration, and vibration intensity. Pressure sensors, current sensors, and temperature sensors are installed on key components such as the air conditioning compressor, condenser, evaporator, and fan motor to collect key parameters such as compressor current / voltage, condenser pressure, evaporator temperature, and fan motor speed.

[0045] Based on environmental data, the system checks at preset intervals, such as every 10 minutes, whether the performance of the mining air conditioner has deteriorated. When performance deteriorates, it further identifies key abnormal parameters and triggers different levels of warnings based on the severity of the deterioration and the type of parameter abnormality. For example, a yellow warning is issued for slight performance degradation, reminding relevant personnel to pay attention to the monitoring of key parameters; if a key parameter is abnormal, a red warning is triggered, requiring the location of the cause of the fault based on the abnormal parameter. The specific methods for determining whether air conditioner performance has deteriorated and for identifying abnormal parameters will be explained later.

[0046] Step S2: If an anomaly occurs, record the time from the last time the mining truck air conditioner started normal operation to the time when the anomaly occurred, as well as the key parameters of the anomaly, which are defined as the first time and the target parameter, respectively. Based on the first time, extract multiple historical reference time periods from the preset historical fault data table.

[0047] Specifically, the time span (e.g., time T) from the last normal operation of the mining air conditioner (without performance degradation) to the detection of the abnormality is obtained, as well as the key parameters of the abnormality, such as the compressor current. The abnormal compressor current is a typical multi-cause-one-effect parameter, which may be caused by a variety of faults, such as refrigerant leakage, reduced compressor efficiency, and short circuit in the motor windings. How to determine the real cause of the abnormal compressor current will be explained later.

[0048] The database stores historical fault data tables containing various fault diagnosis information, such as Figure 2 As shown, the data table includes the event ID, the entire historical cycle from the end of the last maintenance to the occurrence of the anomaly and then to the maintenance, the air conditioning component that failed, the cause of the failure, and the key parameters of the anomaly. Based on multiple historical cycles, multiple historical reference time periods are obtained from the historical fault data table. The specific extraction method will be explained later. The historical reference time period is the time from the detection of the abnormal parameter to the end of maintenance in each historical cycle. By referring to the cause of the failure and the running time of the historical reference time period, the cause of the failure when the abnormal parameter was detected in this round can be further analyzed, thereby improving the accuracy of fault location.

[0049] Step S3: Obtain the causes of failure of the mining truck air conditioner in each historical reference period, calculate the first probability of the target parameter becoming abnormal under each failure cause, and the second probability caused by each failure cause when the components of the mining truck air conditioner fail.

[0050] Specifically, the causes of failures for each reference period are extracted from the historical failure table, such as refrigerant leakage, reduced compressor efficiency, and short circuit in the motor windings. Based on historical data, the first probability of target parameter anomaly under each failure cause is calculated, as well as the second probability caused by each failure cause when a component fails in the mining truck air conditioner. The calculation process will be explained in detail later. A single parameter anomaly may be caused by multiple reasons. Through multi-dimensional correlation analysis, misjudgment based on a single indicator is avoided, and the robustness of diagnosis is improved.

[0051] Step S4: Calculate the third probability of each fault cause when the target parameter is abnormal, based on the first and second probabilities. Determine the fault maintenance sequence of the mining air conditioner based on the third probability and calculate the optimal maintenance time.

[0052] Specifically, the third probability of a mining truck air conditioner failing for each fault cause is calculated based on the first and second probabilities; the calculation method will be explained later. Then, because mining truck air conditioners operate under high temperatures and high dust concentrations for extended periods, environmental parameters have a significant impact on the probability of failure. Therefore, the third probability is adjusted based on environmental data; the specific process will be explained later.

[0053] The causes of failures are sorted based on the adjusted third probability to generate a maintenance priority queue. For example, if the third probability of a certain failure cause is >85%, a high-priority warning is triggered directly. If the probabilities of multiple causes are similar (e.g., the difference is <5%), it is recommended to check related components in parallel. Based on historical data and the degree of anomaly of the current parameters, the optimal maintenance time is calculated.

[0054] In one specific embodiment, determining whether the performance of the mining truck air conditioner has degraded at preset time intervals specifically includes the following steps:

[0055] Environmental data includes the initial temperature and current temperature inside the cab of the mining truck. The mining truck air conditioner has multiple operating modes, each corresponding to an initial temperature range, a target temperature, and a baseline adjustment time. The baseline adjustment time is the average time required to adjust from the historical initial temperature inside the cab to the corresponding target temperature under normal operating conditions of the mining truck air conditioner. The current operating mode of the mining truck air conditioner is obtained based on the initial temperature.

[0056] Determine whether the current temperature has reached the target temperature corresponding to the current operating mode. If it has, record the time required to adjust from the starting temperature to the target temperature, which is defined as the current adjustment time. Calculate the first deviation degree of the mining truck air conditioner based on the current adjustment time and the benchmark adjustment time. If the first deviation degree is greater than the first threshold, it is determined that the mining truck air conditioner has deteriorated.

[0057] Specifically, two environmental data points are collected: the initial temperature of the cab (i.e., the temperature at which the air conditioning starts operating) and the current temperature. The mining truck's air conditioning system has multiple operating modes, such as cooling, heating, or dehumidification. Each mode corresponds to a different initial temperature range, target temperature, and baseline adjustment time. The baseline adjustment time is the average time required for the mining truck's air conditioning to adjust from the initial temperature of the cab to the target temperature under normal operating conditions. Based on an initial temperature, for example, the current temperature being 42℃, and requiring cooling, the current operating mode of the mining truck's air conditioning system is cooling mode.

[0058] The system determines whether the current temperature has reached the target temperature, which is the optimal operating temperature set in cooling mode, such as 22℃. When the current temperature drops to the target temperature, the time required for the drop is recorded and defined as the current adjustment time. The difference between the current adjustment time and the baseline adjustment time obtained from historical data under the same operating mode is calculated to obtain the first deviation level. If the first deviation level is greater than the first threshold (depending on the actual situation), it indicates that the mining air conditioner has deteriorated. The degree of performance degradation is judged based on the first threshold. An early warning is triggered when the performance drops by 10%. Each key parameter is detected to avoid wasting time by detecting the air conditioner only after it has malfunctioned.

[0059] In one specific embodiment, determining whether a key parameter has become abnormal includes the following steps:

[0060] A normal prediction model is created for each sensor as the target sensor. The historical detection values ​​of key parameters of all other sensors under normal conditions are used as input data for the normal prediction model. The normal prediction model learns the correlation between multiple sensors under normal conditions and outputs the predicted value of the target sensor under normal conditions. The second deviation degree between the current detection value and the predicted value of the key parameter of each target sensor is calculated. It is determined whether there is a second deviation degree greater than the second threshold. If there is no second deviation degree, it means that the sensor of the mining air conditioner is not abnormal. If there is a second deviation degree greater than the second threshold, it is determined that the key parameter corresponding to the second deviation degree is abnormal.

[0061] Specifically, each sensor is treated as a target sensor, and a normal prediction model is created for each target sensor. Assuming the target sensor is the compressor, multi-sensor data under normal conditions is extracted from the historical database of mining air conditioners, including historical detection values ​​of compressor current and other sensor data (such as vibration, temperature, pressure, and dust). The data from each sensor is normalized, and a normal prediction model is built using a deep model such as a two-branch Transformer. Historical data from all other sensors under normal conditions are input into the normal prediction model of the target sensor. The normal prediction model learns the correlation between key parameters of the target sensor and other sensor parameters under normal conditions. Real-time data from other sensors (e.g., vibration = 12.5kHz, temperature = 40℃, pressure = 2.1MPa) is obtained and input into the trained model to obtain the predicted value of the target sensor (e.g., predicted current = 12A). The second degree of deviation is calculated based on the following formula. Assuming the actual current is 15A, the second deviation level is 25%. If the deviation level is greater than the second threshold of 20%, it is set as abnormal. Under high temperature conditions, the second threshold can be relaxed to 25%. Mining truck air conditioners are in an environment with high temperature (>35℃), high dust (PM10>200μg / m³), and high vibration (>4mm / s) for a long time. High temperature will cause the compressor current reference value to deviate (such as the current naturally increasing with the temperature). This application establishes a normal prediction model by using a deep model such as a dual-branch Transformer to capture the spatial correlation between different sensor parameters, such as the hysteresis correlation between condenser pressure and compressor current. In extreme working conditions such as mining truck air conditioners, the accuracy of fault warning is improved compared with single sensor prediction.

[0062] In one specific embodiment, extracting multiple historical reference time periods from a preset historical fault data table based on a first-time event specifically includes the following steps:

[0063] The historical fault data table includes the historical cycle of the mining air conditioner from the end of the previous maintenance cycle to the occurrence of the abnormality and then to the end of the current maintenance cycle, the name of the component that failed, the cause of the failure, and the key parameters of the abnormality at the time of the failure. For each historical cycle, the equivalent starting point is calculated. The equivalent starting point is the time point after the first time after the end time of the previous maintenance cycle. The time period from the equivalent starting point to the end of the current maintenance cycle is defined as the historical reference period.

[0064] Specifically, such as Figure 3The diagram shows a historical parameter time period. The horizontal axis represents time, and the vertical axis represents whether the air conditioner is currently running. 1 indicates that the air conditioner is running, and 0 indicates that the air conditioner has malfunctioned. Assume that at time t11, an abnormality in the compressor current is detected. t10-t11 is the time from the end of the last maintenance of the mining air conditioner to the first time the abnormality is detected. t1-t4, t4-t7, and t7-t10 are respectively historical period 1 (i.e., the time period from the end of the last maintenance to normal operation and then to the detection of the abnormality and maintenance), historical period 2, and historical period 3. Assume that the key parameter for the abnormality in all three historical periods is the compressor current.

[0065] Based on the first time, the historical period is divided. For example, in historical period 1, t1-t3 represents the operation phase of the mining truck air conditioner, t3-t4 is the time when the air conditioner malfunctions and needs maintenance. After maintenance, the mining truck air conditioner starts running again. t1 represents the end time of the previous maintenance cycle, and t1-t2 is the running time of the mining truck air conditioner under the same first time. That is, the time point t2 is the equivalent time point of historical period 1. t2-t4 is defined as the historical reference period of historical period 1. Here, t2-t4 includes t2-t3, which is the stage when the air conditioner's key parameters are abnormal but the air conditioner is still running, and t3-t4, which is the stage when the air conditioner malfunctions and needs maintenance. The cause of the malfunction in t2-t3 is also recorded in the historical malfunction data table. This cause of malfunction has important reference value for analyzing the cause of abnormal compressor current in the current period. Similarly, the historical reference periods for historical periods 2 and 3 are t5-t7 and t8-t10, respectively. The above are just examples of three historical periods. All historical reference periods within the past 6 months can be obtained. This application finds similar historical operating cycles by matching timelines, extracts their fault records, analyzes the relationship between abnormal parameters and fault causes, and provides a basis for diagnosing current anomalies.

[0066] In one specific embodiment, calculating the first probability of the target parameter becoming abnormal under each fault cause condition specifically includes the following steps:

[0067] Obtain the first number of times each fault cause causes a mining truck air conditioner fault in the historical fault data table, and the second number of times the corresponding key parameter of the abnormality occurs when the fault cause occurs, with the target parameter being the ratio of the second number to the first number. Define the first probability as the ratio of the second number to the first number.

[0068] Specifically, the first probability refers to the conditional probability that the target parameter (such as compressor current) will become abnormal given that the fault cause i has occurred. It is calculated based on Formula 1, which states that the first probability of the target parameter becoming abnormal under fault cause i is: , This refers to the number of times that fault cause i in the historical fault data table caused the mining truck air conditioner to fail. The key parameter corresponding to the abnormality when fault cause i occurs is the second time of the compressor current. Assuming the target parameter is the compressor current, when the compressor current is abnormal, the fault causes include refrigerant leakage, reduced compressor efficiency, and short circuit of the motor winding.

[0069] Assuming that the number of times the mining air conditioner malfunctions due to refrigerant leakage is 50, and the number of times the compressor current becomes abnormal when refrigerant leakage occurs is 30, then under the condition of refrigerant leakage as the cause of the malfunction, the first probability of the compressor current becoming abnormal is: 0.6; Similarly, the number of times the reduced compressor efficiency caused the mining air conditioner to malfunction is 30, and the number of times the compressor current becomes abnormal when the compressor efficiency is reduced is 15. Therefore, under the condition that the compressor efficiency is reduced as the cause of the malfunction, the first probability of the compressor current becoming abnormal is: 0.5; The number of times the motor winding short circuit caused the mining truck air conditioner to malfunction was 20, and the number of times the compressor current became abnormal when the motor winding short circuit occurred was 5. Therefore, under the condition that the motor winding short circuit was the cause of the malfunction, the first probability of the compressor current becoming abnormal is: 0.25.

[0070] In one specific embodiment, calculating the second probability caused by each fault cause when a component of the mining air conditioner fails specifically includes the following steps:

[0071] Obtain the third number of times each component fails in the historical fault data table and the fourth number of times each component fails due to the corresponding fault cause. Define the ratio of the fourth number to the third number as the second probability.

[0072] Specifically, based on Formula 2, when component j of the mining air conditioner fails, the second probability of i caused by each failure cause is calculated. Formula 2 is as follows: , Let j be the third number of times component j has failed. Let j be the fourth number of times component j fails due to the corresponding fault cause i. Assuming the air conditioner compressor fails 100 times, and refrigerant leakage, reduced compressor efficiency, and motor winding short circuit cause 50, 30, and 20 times respectively, then when the air conditioner compressor fails, the second probability caused by refrigerant leakage is... The second probability caused by reduced compressor efficiency The second probability caused by a short circuit in the motor windings .

[0073] In one specific embodiment, calculating the third probability of each fault cause occurring when the target parameter is abnormal, based on the first and second probabilities, specifically includes the following steps:

[0074] The sum of the products of the first and second probabilities corresponding to all fault causes is defined as the total probability of the target parameter becoming abnormal. The ratio of the product of the first and second probabilities corresponding to each fault cause to the total probability is defined as the third probability of each fault cause occurring when the target parameter becomes abnormal.

[0075] Specifically, the total probability of the target parameter becoming abnormal is calculated based on Formula 3, which is: For example, the probability of an abnormal compressor current is: 0.6 + 0.5 +0.25 Based on Formula 4, the third probability of each fault cause occurring when the target parameter is abnormal is calculated. Formula 4: Therefore, when an abnormality is detected in the target parameter, i.e., the compressor current, the third probability of refrigerant leakage, reduced compressor efficiency, and short circuit in the motor windings are 0.6, 0.3, and 0.1, respectively.

[0076] Adjusting the third probability based on environmental data includes the following steps:

[0077] Obtain a historical fault environment data table, which includes multiple historical environment data and the corresponding relationships of fault causes. Build a logistic regression model based on the historical fault environment data. Calculate the sensitivity coefficient for each environmental parameter in the logistic regression model based on maximum likelihood estimation. Obtain the current environment data. Calculate the distance difference between the current environment data and the historical environment data based on the sensitivity coefficient. Calculate the environment similarity based on the distance difference and introduce a time influence coefficient. Calculate the comprehensive probability influence value of historical fault environment data on each fault cause based on the time influence coefficient of historical environment data and the environment similarity. Adjust the third probability based on the comprehensive probability influence value.

[0078] Specifically, the logistic regression model is as follows: ,in, The probability of the cause of the failure occurring. For constant terms, , ...... Environmental parameters, such as temperature and dust concentration, , ...... These are the logistic regression coefficients, which are also the sensitivity coefficients for each environmental parameter. The coefficients are solved using maximum likelihood estimation. , , ...... Solution method: Since the log-likelihood function is a convex function, iterative optimization algorithms (such as gradient descent and Newton-Raphson method) can be used to solve for the coefficients corresponding to the maximum log-likelihood. For fault causes such as refrigerant leakage, assuming two environmental parameters exist: temperature and dust concentration, after calculation using a logistic regression model, the constant term... =-5, temperature sensitivity coefficient =0.15, sensitivity coefficient of dust concentration =0.10, =0.15 means that for every 1°C increase in temperature, the probability of a failure increases by 0.15. =0.10 means that for every 1 μg / m³ increase in dust concentration, the probability of a fault increases by 0.10.

[0079] The distance difference between the current environment and the historical environment is calculated based on Formula 5, which is: ,in, For the nth environmental sample in historical data One environmental parameter, This represents the k-th environmental parameter value in the current environment. Let be the sensitivity coefficient of the k-th environmental parameter. Assume the current environmental temperature is 50℃ and the dust concentration is 130μg / m³, while the historical environmental temperature was 45℃ and the dust concentration was 120μg / m³. Substitute these values ​​into the calculation to determine the distance difference with the n-th environmental sample. =0.75+1.0=1.75.

[0080] Environmental similarity is calculated based on Formula 6. Formula 6 is: σ is the standard deviation parameter, which controls the sensitivity of similarity. Assuming σ = 10, we substitute it into the calculation of environmental similarity. The value is approximately 0.969, indicating that the current environment is very similar to the historical environment in the nth environmental sample.

[0081] Calculate the time influence coefficient based on Formula 7. Formula 7 is: ,in, The time-related factor represents the impact of equipment aging on fault prediction. It is the decay factor, usually chosen as 0.95 (meaning that the influence of historical data on the prediction will decrease by 5% over time). For the current time, Let the timestamp corresponding to the historical data n be substituted into the formula to calculate the decay factor. Assuming that the time difference between the current time and the historical time is 1 year, the decay factor corresponding to the 1-year time difference is 0.95, which means that the influence weight of the historical data is 95% of that of the current data.

[0082] Calculate the overall probability impact value based on Formula 8. ,in, ,in, Let be the probability of failure i occurring under historical environment n (usually the probability of each failure cause obtained from historical data).

[0083] Assume the current ambient temperature is 45℃ and the dust concentration is 120μg / m³.

[0084] The probability of failure (assuming refrigerant leakage) occurring in historical environment 1 (temperature = 45℃, dust concentration = 120μg / m³) is 0.391, with an environmental similarity of 0.391. Time Influence Coefficient The values ​​are 1 and 0.377, respectively.

[0085] The probability of refrigerant leakage in historical environment 2 (temperature = 50℃, dust concentration = 130μg / m³) is 0.55, with an environmental similarity of 0.55. Time Influence Coefficient The values ​​are 0.85 and 0.377 respectively; the failure probability corresponding to historical environment 3 (temperature = 35℃, dust concentration = 90μg / m³) is 0.7, and the environmental similarity is... Time Influence Coefficient The values ​​are 0.75 and 0.044 respectively (because the historical data is very old, the impact is very small).

[0086] The weighted probability of historical environment 1 is then 1. 0.377 0.391 = 0.147; Weighted probability of historical environment 2 = 0.85 0.377 0.55 = 0.176; Weighted probability of historical environment 3 = 0.75 0.004 0.70 = 0.0021, therefore the combined probability impact value = 0.147 + 0.176 + 0.0021 = 0.3251. The third probability is adjusted based on the combined probability impact value, and the product of the combined probability impact value and the third probability is taken as the adjusted third probability.

[0087] In actual operation, equipment often faces multiple uncertain environmental factors (such as temperature fluctuations, humidity changes, dust accumulation, etc.) that may negatively affect the long-term performance of the equipment. By adjusting the third probability, the impact of the environment on equipment failure can be dynamically strengthened or weakened, reducing misjudgments and omissions caused by changes in environmental factors.

[0088] In one specific embodiment, determining the fault maintenance sequence of the mining air conditioner based on a third probability includes the following steps:

[0089] Based on the historical fault data table, the first distribution function corresponding to each fault cause is obtained. The first distribution function is the distribution function of the historical probability of the mining air conditioner failing before the key parameters are abnormal under the influence of the fault cause as a function of running time. The first time is substituted into the first distribution function to obtain the first function value. The product of the third probability corresponding to each fault cause and the first function value is defined as the fault probability of the mining air conditioner failing due to the fault cause in the first time. Fault maintenance is performed on the mining air conditioner based on each fault cause in descending order of fault probability.

[0090] Specifically, different failure causes have different failure modes, therefore the first distribution function for different failure causes is usually also different. For example, the first distribution function for refrigerant leakage is a Weibull distribution because refrigerant leakage is usually a slow process, possibly caused by aging of seals, fatigue cracks due to vibration, or corrosion, etc. The leakage rate may gradually increase over time. The Weibull distribution is often used to describe failure modes such as fatigue and wear. Therefore, the formula for the first distribution function of refrigerant leakage is: ,in This is a shape parameter, which may be less than 1 (early failure) or greater than 1 (wear-out failure), for example... =2 indicates that the failure rate increases with time, η is the scale parameter (characteristic lifetime), for example, η=8000 hours (meaning that about 63.2% of the units fail at 8000 hours), and t is the running time, for example, the first time.

[0091] The first distribution function for compressor efficiency reduction is a log-normal distribution because compressor efficiency reduction can be due to internal wear (such as piston ring wear, increased cylinder clearance, etc.), leading to a decrease in the compression ratio. This wear is usually gradual and continuous, and the log-normal distribution is often used to describe the wear process because wear is often related to the logarithm of time. The Weibull distribution is also applicable (β>1). The first distribution function for compressor efficiency reduction is... , where Φ is the cumulative distribution function of the standard normal distribution, and μ and σ are parameters of the log-normal distribution. For example, μ = 8.5 (corresponding to a median lifetime of approximately 5000 hours), σ = 0.5.

[0092] The first distribution function of a short circuit in a motor winding is either a Weibull distribution or an exponential distribution. However, insulation aging is usually time-dependent, so the Weibull distribution (β>1) is more common. A winding short circuit may be caused by insulation failure due to aging of the insulation material (heat, electrical stress, moisture, etc.), leading to a short circuit. This failure may be sudden (e.g., insulation breakdown). The first distribution function is: Where β=1.5, =10,000 hours. The distribution function for each of the above fault causes is obtained by fitting historical fault data. For example: collect the occurrence time of 50 refrigerant leak events, fit the β and η of the Weibull distribution, and when an abnormal compressor current is detected, combine the current running time, i.e., the first time T, and substitute the first time into the first distribution function corresponding to each fault cause to calculate the first function value. The product of the third probability corresponding to each cause of failure and the first function value is... Defined as the probability of a mining truck air conditioner malfunctioning due to fault cause i within the first time period.

[0093] In one specific embodiment, calculating the optimal maintenance time includes the following steps:

[0094] The sum of the failure probabilities corresponding to all failure causes is defined as the total probability of the mining truck air conditioner failing in the first time period. As the mining truck air conditioner runs, it is judged at preset time intervals whether the total probability is greater than the third threshold. If it is, the running time of the mining truck air conditioner at this time is defined as the second time point, and the time period between the first time point and the second time point is defined as the optimal maintenance time.

[0095] Specifically, the failure probabilities corresponding to each failure cause are summed to obtain the total probability of the mining truck air conditioner failing in the first time period. After detecting a target parameter such as an abnormal compressor current, the total probability is checked at preset intervals to see if it is greater than a third threshold. Since the total probability changes with the running time, the running time when the total probability is greater than the third threshold is defined as the second time point. After this time point, it means that the mining truck air conditioner will fail. Therefore, the time period between the first time point and the second time point is defined as the optimal maintenance time, that is, maintenance is carried out before a failure occurs, thereby improving the operating efficiency of the mining truck air conditioner.

[0096] The above describes a data analysis-based fault diagnosis method for mining truck air conditioners in the embodiments of this application. The following describes a data analysis-based fault diagnosis system for mining truck air conditioners in the embodiments of this application. Please refer to [link / reference]. Figure 4 One embodiment of a data analysis-based fault diagnosis system for mining truck air conditioners in this application includes:

[0097] The monitoring module is used to deploy multiple sensors on the relevant components of the mining truck air conditioner to acquire environmental data of the mining truck and the current detection values ​​of key parameters in the sensors. It determines whether the performance of the mining truck air conditioner has deteriorated at preset intervals, and if so, determines whether the key parameters have become abnormal.

[0098] The extraction module is used to record the time from the most recent start of normal operation of the mining truck air conditioner to the occurrence of the abnormality, as well as the key parameters of the abnormality, if an abnormality occurs. These are defined as the first time and the target parameter, respectively. Based on the first time, multiple historical reference time periods are extracted from the preset historical fault data table.

[0099] The calculation module obtains the causes of failures in the mining truck air conditioner within each historical reference period, calculates the first probability of the target parameter becoming abnormal under each failure cause, and the second probability caused by each failure cause when the components of the mining truck air conditioner fail.

[0100] The maintenance module is used to calculate the third probability of each fault cause when the target parameter is abnormal, based on the first and second probabilities. Based on the third probability, it determines the fault maintenance sequence of the mining air conditioner and calculates the optimal maintenance time.

[0101] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0102] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0103] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A data analysis-based method for diagnosing faults in mining truck air conditioners, characterized in that, The method includes: Step S1: Deploy multiple sensors on the relevant components of the mining truck air conditioner to obtain environmental data of the mining truck and the current detection values ​​of key parameters in the sensors. Determine whether the performance of the mining truck air conditioner has degraded at preset intervals. If so, determine whether the key parameters have become abnormal. The process of determining whether the performance of the mining truck air conditioner has degraded at preset intervals includes: the environmental data including the initial temperature and current temperature inside the mining truck cab; the mining truck air conditioner has multiple operating modes, each corresponding to an initial temperature range, a target temperature, and a baseline adjustment time; the baseline adjustment time is the average time required to adjust from the historical initial temperature inside the cab to the corresponding target temperature under normal operating conditions; the current operating mode of the mining truck air conditioner is obtained based on the initial temperature; it is determined whether the current temperature has reached the target temperature corresponding to the current operating mode; if so, the time required to adjust from the initial temperature to the target temperature is recorded and defined as the current adjustment time; a first deviation degree of the mining truck air conditioner is calculated based on the current adjustment time and the baseline adjustment time; if the first deviation degree is greater than a first threshold, it is determined that the mining truck air conditioner has degraded. Step S2: If an anomaly occurs, record the time from the most recent start of normal operation of the mining truck air conditioner to the occurrence of the anomaly, as well as the key parameters of the anomaly, which are defined as the first time and the target parameter, respectively. Based on the first time, extract multiple historical reference time periods from the preset historical fault data table. Step S3: Obtain the causes of failure of the mining truck air conditioner in each historical reference period, calculate the first probability of the target parameter becoming abnormal under each failure cause, and the second probability caused by each failure cause when the components of the mining truck air conditioner fail. Step S4: Calculate the third probability of each fault cause when the target parameter is abnormal based on the first probability and the second probability, determine the fault maintenance sequence of the mining air conditioner based on the third probability, and calculate the optimal maintenance time.

2. The method according to claim 1, characterized in that, Determining whether the key parameters are abnormal includes: A normal prediction model is created for each sensor as the target sensor. The historical detection values ​​of key parameters of all other sensors under normal conditions are used as input data for the normal prediction model. The normal prediction model learns the correlation between multiple sensors under normal conditions and outputs the predicted value of the target sensor under normal conditions. The second deviation degree between the current detection value and the predicted value of the key parameter of each target sensor is calculated. It is determined whether there is a second deviation degree greater than a second threshold. If there is no deviation degree, it means that the sensor of the mining air conditioner is not abnormal. If there is a deviation degree, it is determined that the key parameter corresponding to the second deviation degree greater than the second threshold is abnormal.

3. The method according to claim 1, characterized in that, Based on the first time, multiple historical reference time periods are extracted from a preset historical fault data table, including: The historical fault data table includes the historical cycle of the mining air conditioner from the end of the previous maintenance cycle to the occurrence of an anomaly and then to the end of the current maintenance cycle, the name of the component that failed, the cause of the failure, and the key parameters of the anomaly at the time of the failure. For each historical cycle, an equivalent starting point is calculated. The equivalent starting point is the time point after adding the first time to the end time of the previous maintenance cycle. The time period from the equivalent starting point to the end of the current maintenance cycle is defined as the historical reference period.

4. The method according to claim 3, characterized in that, Calculating the first probability of the target parameter becoming abnormal under each fault cause condition includes: The first number of times each fault cause caused the mining truck air conditioner to fail in the historical fault data table and the second number of times the key parameter of the abnormality corresponding to the fault cause occurred are obtained as the target parameter. The ratio of the second number to the first number is defined as the first probability.

5. The method according to claim 4, characterized in that, Calculate the second probability caused by each cause of component failure in the mining air conditioner, including: Obtain the third number of times each component fails and the fourth number of times each component fails due to the corresponding fault cause from the historical fault data table, and define the ratio of the fourth number to the third number as the second probability.

6. The method according to claim 5, characterized in that, Based on the first probability and the second probability, a third probability of each fault cause occurring when the target parameter becomes abnormal is calculated, including: The sum of the products of the first probability and the second probability corresponding to all fault causes is defined as the total probability of the target parameter becoming abnormal. The ratio of the product of the first probability and the second probability corresponding to each fault cause to the total probability is defined as the third probability of each fault cause occurring when the target parameter becomes abnormal.

7. The method according to claim 1, characterized in that, Determining the fault maintenance sequence of the mining truck air conditioner based on the third probability includes: Based on the historical fault data table, obtain the first distribution function corresponding to each fault cause. The first distribution function is the distribution function of the historical probability of the mining air conditioner failing before the key parameters are abnormal under the action of the fault cause, as a function of the running time. Substitute the first time into the first distribution function to obtain the first function value. The product of the third probability corresponding to each fault cause and the first function value is defined as the fault probability of the mining air conditioner failing due to the fault cause within the first time period. Perform fault maintenance on the mining air conditioner based on each fault cause in descending order of the fault probability.

8. The method according to claim 7, characterized in that, Calculating the optimal maintenance time includes: The sum of the failure probabilities corresponding to all failure causes is defined as the total probability of the mining truck air conditioner failing within the first time period. As the mining truck air conditioner runs, it is determined at preset intervals whether the total probability is greater than a third threshold. If it is, the running time of the mining truck air conditioner at this time is defined as the second time point, and the time period between the first time point and the second time point is defined as the optimal maintenance time.

9. A data analysis-based fault diagnosis system for mining truck air conditioners, used to implement the data analysis-based fault diagnosis method for mining truck air conditioners as described in any one of claims 1-8, characterized in that, The system includes: The monitoring module is used to deploy multiple sensors on the relevant components of the mining truck air conditioner to acquire environmental data of the mining truck and the current detection values ​​of key parameters in the sensors. It determines whether the performance of the mining truck air conditioner has deteriorated at preset intervals, and if so, determines whether the key parameters have become abnormal. The extraction module is used to record the time from the most recent start of normal operation of the mining truck air conditioner to the occurrence of the abnormality, as well as the key parameters of the abnormality, if an abnormality occurs. These are defined as the first time and the target parameter, respectively. Based on the first time, multiple historical reference time periods are extracted from a preset historical fault data table. The calculation module obtains the causes of failure of the mining truck air conditioner in each historical reference period, calculates the first probability of the target parameter becoming abnormal under each failure cause, and the second probability caused by each failure cause when the components of the mining truck air conditioner fail. The maintenance module is used to calculate the third probability of each fault cause when the target parameter is abnormal, based on the first probability and the second probability, determine the fault maintenance sequence of the mining air conditioner based on the third probability, and calculate the optimal maintenance time.

Images