A life prediction system and air conditioning system

Abstract

The application discloses a life prediction system and an air conditioning system, and relates to the technical field of life prediction systems, and can accurately predict the life of a device to be predicted according to environmental parameters and operation parameters of the device to be predicted, so that the device to be predicted can be maintained and replaced in advance, the service life of the entire system can be effectively prolonged, and the risk of failure and resource waste can be reduced.

Description

Technical Field

[0001] This invention relates to the field of refrigeration, and in particular to a life prediction system and an air conditioning system. Background Technology

[0002] During long-term use, the internal components of an air conditioning system gradually age due to various factors, leading to performance degradation or even failure. This is especially true for vulnerable parts such as seals and compressor lights, whose lifespan directly impacts the stability and reliability of the entire air conditioning system.

[0003] Traditional maintenance methods typically rely on periodic replacement or repair, but this approach is often imprecise and may lead to wasted resources or increased risk of failure.

[0004] Therefore, how to design a life prediction system and air conditioning system that can predict the remaining life of each component in the air conditioning system so as to carry out maintenance and replacement in advance, reduce the risk of failure and waste of resources, is a technical problem that the industry urgently needs to solve. Summary of the Invention

[0005] In view of the problem that traditional maintenance methods in the prior art rely on periodic replacement or repair, which may lead to waste of resources or increased risk of failure, this invention proposes a life prediction system and an air conditioning system.

[0006] The technical solution of this invention is to propose a lifetime prediction system, comprising:

[0007] The data acquisition module is used to collect the operating data of the device to be predicted;

[0008] The data analysis module is used to calculate the life loss of the device to be predicted based on the operating data.

[0009] A lifespan prediction module is used to estimate the remaining lifespan of the device to be predicted based on the lifespan loss.

[0010] Furthermore, the data acquisition module includes a first acquisition device for acquiring environmental parameters of the device to be predicted, and a second acquisition device for acquiring operating parameters of the device to be predicted.

[0011] The sampling parameters of the second acquisition device are set according to the environmental parameters.

[0012] Furthermore, when acquiring the operational data, the data acquisition module performs the following actions:

[0013] The weights of the operating parameters are set according to the environmental parameters.

[0014] Furthermore, when acquiring the operational data, the data acquisition module also performs the following actions:

[0015] Determine whether the currently collected running data exceeds the threshold range;

[0016] If so, the currently collected running data will be corrected and repaired.

[0017] Furthermore, the calculation model for the lifespan loss of the device to be predicted by the data analysis module is as follows:

[0018] y=β0+β1x1+β2x2+……+βpxp+ε

[0019] Where y is the lifespan loss, x1, x2, ..., xp are the operating data, β0, β1, ..., βp are the proportional coefficients, and ε is the error value.

[0020] Furthermore, the calculation model used by the lifetime prediction module to estimate the remaining lifetime of the device to be predicted is as follows:

[0021] F=Ny;

[0022] Where F is the remaining lifetime, N is the ideal lifetime of the device to be predicted, and y is the lifetime loss.

[0023] Furthermore, the sampling parameters include at least sampling time, sampling frequency, and anti-interference capability.

[0024] Furthermore, the lifetime prediction system also includes an adaptive learning module, which is used to train and optimize the calculation model for calculating lifetime loss and the calculation model for calculating remaining lifetime using the batch gradient descent method.

[0025] Furthermore, the training model for the batch gradient descent method is as follows:

[0026]

[0027] in, , For a single selected sample, Let be the gradient of the loss function for a single sample, and m be the number of samples used for training. Let be the hypothesis function in the batch gradient descent method. This is the weight vector.

[0028] The present invention also proposes an air conditioning system having the above-mentioned life prediction system.

[0029] Compared with the prior art, the present invention has at least the following beneficial effects:

[0030] The life prediction system proposed in this invention has a data acquisition module, a data analysis module, and a life prediction module. It can collect environmental parameters and operating parameters of the device to be predicted, calculate the life loss of the device to be predicted, and finally estimate the remaining life of the device to be predicted. When this invention is applied to an air conditioning system, it can be used to calculate the life loss of each component in the air conditioning system and predict the remaining life of each component accordingly, so as to carry out maintenance and replacement in advance and effectively extend the service life of the air conditioning system.

[0031] Furthermore, the life prediction system proposed in this invention also has an adaptive learning module, which can optimize the calculation models in the data analysis module and the life prediction module based on a large amount of operational data, thereby creating a calculation model that can accurately predict the remaining life of the device to be predicted, thus improving the accuracy of the prediction of the remaining life of the device to be predicted in this invention. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a schematic diagram of the overall modular structure of the present invention;

[0034] Figure 2 This is a flowchart illustrating the overall operation of the present invention.

[0035] Figure 3 This is a block diagram of the data acquisition module in this invention. Detailed Implementation

[0036] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0037] Therefore, a feature pointed out in this specification is used to describe one feature of one embodiment of the invention, and does not imply that every embodiment of the invention must have the described feature. Furthermore, it should be noted that this specification describes many features. Although certain features may be combined to illustrate possible system designs, these features may also be used in other combinations not explicitly stated. Therefore, unless otherwise stated, the described combinations are not intended to be limiting.

[0038] The principles and structure of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0039] During long-term use, the internal components of an air conditioning system gradually age due to various factors, leading to performance degradation or even failure. This is especially true for vulnerable components such as seals and compressors, whose lifespan directly impacts the stability and reliability of the entire air conditioning system.

[0040] Current solutions to these problems rely on periodic replacement or maintenance, but this approach is often imprecise and may lead to wasted resources or increased risk of failure.

[0041] Based on the problems existing in the existing technical solutions, the present invention proposes a life prediction system, which aims to predict the remaining life of the device to be predicted. The life prediction system of the present invention can be applied to the air conditioning system to predict the remaining life of each component in the air conditioning system, so that users can perform early maintenance and replacement of components, thereby extending the service life of the entire air conditioning system and reducing the risk of failure and resource risks.

[0042] Please see Figure 1 The lifetime prediction system proposed in this invention includes:

[0043] The data acquisition module is used to collect the operating data of the device to be predicted.

[0044] The data analysis module is used to calculate the life loss of the device to be predicted based on the operating data.

[0045] A lifespan prediction module is used to estimate the remaining lifespan of the device to be predicted based on lifespan loss.

[0046] As can be seen from the above settings, the data acquisition module, data analysis module, and life prediction module in this invention are in a progressive relationship. The data acquisition module can collect the operating data of the device to be predicted, and then transmit the operating data to the data analysis module. The data analysis module can calculate the life loss of the device to be predicted based on this data. After calculating the life loss, the life prediction module can estimate the remaining life of the device to be predicted based on this data.

[0047] The life prediction system of this invention can be applied to air conditioning systems. In this case, the device to be predicted can be a component in the air conditioning system, such as a sealing ring, compressor, fan motor, etc. The life prediction system provided by this invention can accurately predict the remaining life of the sealing ring, compressor, fan motor, etc. Users can maintain and replace the sealing ring, compressor, fan motor, etc. based on the remaining life of the components, thereby avoiding the problem of damage to the entire air conditioning system caused by the wear and tear of the components, effectively extending the service life of the air conditioning system, and reducing the risk of failure and resource risks.

[0048] It should be noted that the air conditioning system is only a preferred embodiment of the present invention. The present invention is not limited to the above examples. For example, the present invention can also be applied to electric vehicles to predict the remaining life of the engine, tires, bearings and other devices in the electric vehicle, so as to perform early maintenance and replacement and avoid the problem of the entire electric vehicle failing due to the damage of the above components.

[0049] Furthermore, since the present invention can predict the remaining lifespan of each component, when an air conditioning system or electric vehicle malfunctions, it can determine the specific faulty component based on the remaining lifespan of each component in the air conditioning system or electric vehicle, thereby improving the accuracy of the user's location of the faulty component.

[0050] In other words, the lifetime prediction system proposed in this invention has the following beneficial effects:

[0051] It can accurately predict the lifespan of the device to be predicted based on the environmental and operating parameters of the device, so as to carry out early maintenance and replacement of the device, effectively extend the service life of the entire system, and reduce the risk of failure and waste of resources.

[0052] Please see Figure 3 In this invention, the data acquisition module includes a first acquisition device for acquiring environmental parameters of the device to be predicted, and a second acquisition device for acquiring operating parameters of the device to be predicted.

[0053] The sampling data of the second acquisition device is set according to environmental parameters.

[0054] Regarding the lifespan loss of predictive devices, the factors affecting their lifespan loss are mainly divided into environmental parameters and operating parameters. For example, for the sealing rings in the air conditioning system, due to the principle of thermal expansion and contraction, the sealing rings are affected differently under different temperatures and light intensities. These temperature and light intensities are environmental parameters, which means that the environmental parameters affect the lifespan loss of the sealing rings.

[0055] Similarly, for the compressor in the air conditioning system, the speed of the compressor rotor is different at different operating frequencies, which leads to different lifespan losses of the compressor. This operating frequency is an operating parameter, which means that the operating parameter affects the lifespan loss of the compressor.

[0056] Therefore, the present invention includes a first acquisition device and a second acquisition device, which are mainly used to acquire the above-mentioned environmental parameters and operating parameters, obtain all operating data that will affect the remaining lifespan of the device to be predicted, and thus accurately calculate the lifespan loss of the device to be predicted for the purpose of estimating the final remaining lifespan.

[0057] The environmental parameters mentioned above include temperature, humidity, light intensity, electromagnetic interference, and weather conditions. The specific environmental parameters to be referenced can be selected according to actual needs, mainly those that will affect the remaining lifespan of the device to be predicted.

[0058] In regions like the Middle East, where sulfur dioxide levels are high, it can corrode the heat exchange fins of air conditioning systems, thus affecting their remaining lifespan. In such cases, the environmental parameters mentioned above should include the solubility or content of sulfur dioxide.

[0059] The specific first acquisition device is set according to actual needs. When the environmental parameters to be acquired are temperature and humidity, the first acquisition device can be a temperature sensor and a humidity sensor.

[0060] When the collected environmental parameter is light intensity, the first collection device mentioned above can be a photothermal sensor;

[0061] The specific configuration of the first data acquisition device is not limited. Any device used to collect environmental parameters should be within the protection scope of this invention.

[0062] Among them, the above-mentioned operating parameters include operating frequency, operating mode, etc. The specific selection is the same as the selection of environmental parameters, mainly selecting operating parameters that will affect the remaining lifespan of the device to be predicted.

[0063] For compressors, their operating frequency affects their remaining lifespan, so the above operating parameters need to include the compressor's operating frequency.

[0064] The specific second acquisition device also needs to be set according to actual needs. For example, when acquiring the voltage of the operating parameter, the above-mentioned second acquisition device can be a voltage sensor.

[0065] For example, when collecting operating parameters such as current, the second acquisition device mentioned above can be a current sensor;

[0066] Even for certain parameters, such as operating frequency, the air conditioning system usually has a record of the operating frequency when the compressor is running. In this case, there is no need to set up a second acquisition device. The operating frequency can be retrieved directly from the air conditioning system.

[0067] The specific configuration of the second acquisition device is not limited. Any device used to acquire operating parameters should be within the protection scope of this invention.

[0068] Furthermore, the above settings also indicate that the sampling parameters of the second acquisition device are set according to environmental parameters;

[0069] The sampling parameters here generally include sampling time, sampling frequency, and anti-interference capability. This is because in some specific environments, such as in places with different electromagnetic interference intensities, if a sensor with the same anti-interference capability is used to collect operating parameters, in places with strong electromagnetic interference, the sensor's poor anti-interference capability may lead to a large error in the collected data.

[0070] In this case, the environmental parameter is the electromagnetic interference intensity, and the sampling parameter is the anti-interference capability. At this time, the anti-interference capability selected by the sensor needs to be selected according to the environmental parameter, that is, the sampling parameter of the second acquisition device mentioned above is set according to the environmental parameter.

[0071] For example, in windy weather, the blades of the outdoor fan of the air conditioning system may rotate in both directions due to the external wind. When the outdoor fan of the air conditioning system is running, the current on the three-phase motor will vary greatly depending on the wind force and whether there is wind. At this time, it is necessary to increase the sampling frequency of the current sensor on the current of the three-phase motor to ensure the accuracy of the sampling of operating parameters.

[0072] In this case, the environmental parameters are the wind speed and intensity, and the sampling parameter is the sampling frequency. At this time, the sampling frequency selected by the current sensor needs to be selected according to the environmental parameters, that is, the sampling parameters of the second acquisition device mentioned above are set according to the environmental parameters.

[0073] In summary, by setting up the first and second sampling devices, this invention can acquire the operating parameters of the device under prediction based on environmental parameters, which can then be used for subsequent calculations of lifespan loss. Furthermore, when sampling the operating parameters, this invention adjusts the sampling parameters according to the environmental parameters, greatly minimizing the influence of environmental parameters on the sampling of operating parameters and ensuring the accuracy of the sampling.

[0074] Furthermore, when acquiring the aforementioned operational data, the data acquisition module in this invention also performs the following actions:

[0075] Set the weights of operating parameters based on environmental parameters;

[0076] This setting is mainly used to improve the accuracy and robustness of the collected operating parameters and ensure the rationality of the collected operating parameters.

[0077] For electric vehicles, when the environmental parameters are sunny and rainy, the electric vehicle is running in the same gear. The life loss of the engine will be much greater in the rain than in the sunny. When calculating the life loss of the engine, only the gear in which the electric vehicle is running is considered, without taking into account the impact of the environmental parameters when the electric vehicle is running in that gear.

[0078] In this invention, the weighting of operating parameters based on environmental parameters is used to avoid errors in lifespan loss calculation caused by the different impacts of the same operating parameter on lifespan loss when the environmental parameters are different.

[0079] Taking the aforementioned electric vehicle as an example, if the weight of the current gear of the electric vehicle under sunny weather is 1, then the weight of the current gear of the electric vehicle under rainy weather can be set to 1.4. In this way, the problem of errors in calculating life loss can be avoided because the same gear has different effects on life loss under different environmental parameters in rainy and sunny weather.

[0080] Furthermore, when acquiring the aforementioned operational data, the data acquisition module in this invention also performs the following actions:

[0081] Determine whether the currently collected running data exceeds the threshold range;

[0082] If so, the currently collected running data will be corrected and repaired.

[0083] This part of the invention is designed to avoid abnormal situations, such as sensor failures, from affecting the collected operational data during the data acquisition process.

[0084] For air conditioning systems, their operating data generally exhibits continuous and complete changes. Therefore, the current operating data can be judged as abnormal based on the operating data collected before and after the previous two collections.

[0085] In this invention, the threshold range can be set using a 15% deviation from the previous data acquisition and a 15% deviation from the subsequent data acquisition to determine whether the currently acquired running data is abnormal. For example, if the previous data acquisition is 100 and the subsequent data acquisition is 200, the range of a 15% deviation from the previous data acquisition is (85, 115), and the range of a 15% deviation from the subsequent data acquisition is (170, 230). In this case, the threshold range is set to (85, 115) and (170, 230). If the current data acquisition is 150, it is determined that the current data acquisition is not within the threshold range and needs to be corrected and repaired.

[0086] Conversely, if the current data collected is 110, which is within the threshold range, then the data can be considered normal.

[0087] Setting the threshold range here to 15% of the deviation of the previous data collection and 15% of the deviation of the next data collection can prevent the air conditioning system's operating conditions from suddenly changing, causing a sharp change in the sampled operating parameters, thus misjudging the abnormality of the current operating parameters.

[0088] With this setting, regardless of whether the change in the air conditioning system's operating condition occurs between the previous and current sampling or between the current and next sampling, as long as the operating parameters do not show any abnormalities, the operating parameters can be guaranteed to fall within the threshold range, thus completely avoiding misjudgments caused by changes in the air conditioning system's operating condition.

[0089] In the above judgment, the method for correcting and repairing the currently collected running data is to use interpolation, smoothing or other data repair methods to ensure the continuity and integrity of the running data.

[0090] In other words, based on the above settings, the present invention can identify and correct abnormal operating data during the acquisition process, thereby ensuring the continuity and integrity of all data.

[0091] Please see Figure 3 The data acquisition steps performed by the aforementioned data acquisition module in this invention include:

[0092] 1. Parameter classification: The collected operational data is divided into environmental parameters and operational parameters, and samples are taken separately.

[0093] 2. Dynamic environmental perception: Monitor environmental data and dynamically adjust the output or calibration parameters of the sensor according to environmental changes to eliminate errors. For example, monitor light intensity in real time and dynamically adjust the gain of the sensor used to collect light intensity under different light intensities to adapt to the environmental parameter collection needs under different light intensities.

[0094] 3. Sampling parameter adjustment: Dynamically adjust the sampling frequency or sensor parameters according to changes in environmental parameters. For example, increase the sampling frequency when the environment changes drastically to capture more detailed information, or use a stronger filter when electromagnetic interference is strong to reduce the impact of noise.

[0095] 4. Set weight ratios to combine data collected from multiple sensors and perform data fusion by weighted averaging, thereby improving the accuracy and robustness of the data;

[0096] 5. Anomaly monitoring and handling: Real-time monitoring of whether the operating data exceeds the threshold range. If so, an alarm is triggered and corresponding remedial measures are taken. For the abnormal operating data detected, differential, smoothing or other data repair methods are used to correct it to ensure the continuity and integrity of the operating data.

[0097] Furthermore, the calculation model for calculating the lifespan loss of the device to be predicted by the data analysis module in this invention is as follows:

[0098] y=β0+β1x1+β2x2+……+βpxp+ε

[0099] Where y is the lifespan loss, x1, x2, ..., xp are the operating data, β0, β1, ..., βp are the proportional coefficients, and ε is the error value.

[0100] Among them, x1, x2...xp can be different operating data. For example, for the sealing ring in the air conditioning system, the above operating data can be sulfur dioxide concentration, temperature, pressure, operating time, etc.

[0101] The operating data x1, x2, ... xp selected here are operating data that affect the lifespan loss of the device to be predicted, such as the sealing ring in the air conditioning system, whose selected operating data are sulfur dioxide concentration, temperature, pressure, operating time, etc.

[0102] β0, β1, ..., βp are generally weighted according to the impact of different operating data on lifespan loss;

[0103] Based on the above calculation model, the present invention can calculate the life loss of the device to be predicted and use it to finally estimate the remaining life of the device to be predicted.

[0104] Furthermore, the calculation model for estimating the remaining lifespan of the device to be predicted by the lifespan prediction module in this invention is as follows:

[0105] F=Ny;

[0106] Where F is the remaining lifetime, N is the ideal lifetime of the device to be predicted, and y is the lifetime loss.

[0107] Here, the ideal service life of each component in the air conditioning system is usually marked when the component leaves the factory and is a known quantity. y is the life loss, which can vary greatly due to the influence of environmental and operating parameters and is an unknown quantity. However, this life loss can be obtained through the life loss calculation model mentioned above and used to calculate the remaining life.

[0108] After using the above-mentioned remaining life calculation model to predict the remaining life of the device to be predicted, the device to be predicted can be replaced according to the remaining life of the device to be predicted, such as when the life prediction system proposed in this invention is applied to the prediction of the remaining life of various components in the air conditioning field.

[0109] If the remaining life of a component is short, it needs to be replaced as soon as possible to avoid the remaining life becoming zero, which would damage the seal and affect the operation of the air conditioning system.

[0110] If a component has a long remaining lifespan, it may not need to be replaced even if it has been used for a long time.

[0111] If the remaining lifespan of a component is zero, it needs to be replaced immediately to prevent damage to the component and its impact on the air conditioning system.

[0112] As can be seen from the above settings, after adopting the life prediction system proposed in this invention, the remaining life information of components can be obtained. Components with short remaining life can be replaced in time to avoid damage to the air conditioning system. Components with long remaining life can be temporarily left unreplaced to save on the maintenance cost of the air conditioning system. Components with zero remaining life need to be replaced immediately to prevent damage to the component and its spread to the air conditioning system.

[0113] Please see Figure 1 Since many factors affect lifetime loss when predicting remaining lifetime, the specific lifetime loss calculation model may also change, which may lead to deviations in the prediction of remaining lifetime in this invention. In order to optimize this problem, the lifetime prediction system proposed in this invention is equipped with an adaptive learning module. This adaptive learning module is used to train and optimize the calculation model for calculating lifetime loss and the calculation model for calculating remaining lifetime according to the batch gradient method.

[0114] By setting up the adaptive learning module, the present invention can continuously optimize the calculation model used to calculate lifetime loss and the calculation model used to calculate remaining lifetime, thereby avoiding the problem of deviation in the prediction of component lifetime loss due to changes in the calculation model.

[0115] The training model for batch gradient descent in this invention is as follows:

[0116]

[0117] in, , For a single selected sample, Let be the gradient of the loss function for a single sample, and m be the number of samples used for training. Let be the hypothesis function in the batch gradient descent method. This is the weight vector.

[0118] Based on the adaptive learning module, the point setting process in this invention is as follows:

[0119] 1. Build a data model using the existing database, which stores runtime parameters and environment parameters;

[0120] 2. Train the computational model using the batch gradient descent method. In each iteration, the entire training set is used to calculate the gradient, and then the parameters of the computational model are updated based on this gradient.

[0121] 3. Then, the parameters in the database are called, and the feature weights are dynamically adjusted using the mathematical idea of ​​random forest (constructing multiple decision trees and aggregating their predictions to make a final prediction; an important feature of random forest is that it can provide an assessment of feature importance). When training multiple decision trees, each decision tree uses a randomly selected subset of features during its construction process, and the importance of each feature is standardized so that its sum is 1.

[0122] 4. Update the parameters of the computational model based on the adjustment results of the adaptive learning module, and repeat the training and evaluation process. Example: Update each parameter using the existing batch gradient descent method, taking environmental parameters such as temperature, humidity, and air corrosion, as well as operating parameters such as compressor current and fan speed. This can be understood as each parameter being determined by a regression function. Then, treat the environmental parameters and operating parameters as decision trees, standardize the importance of each trait, and finally obtain a computational model with the highest credibility. This can be understood as the final computational model being determined by all environmental parameters and operating parameters. The proportion of these parameters varies depending on their importance, and the final sum is 1.

[0123] To improve prediction accuracy, this invention can also introduce robot learning algorithms to predict the lifespan of components by training a model. For example, algorithms such as Support Vector Machine (SVM) and Random Forest can be used to train a model that can accurately predict the lifespan of components based on a large amount of running data.

[0124] Furthermore, the life prediction system provided in this invention also needs to support remote access, allowing maintenance personnel to view real-time life estimation and operating data via the Internet, facilitating remote monitoring and maintenance, and reducing maintenance costs.

[0125] The life prediction system proposed in this invention also includes an alarm module, which automatically sends an alarm to maintenance personnel when it is predicted that the life of the device to be predicted is about to be exhausted, so as to give early warning of potential faults and reduce unexpected downtime.

[0126] Please see Figure 2 The overall workflow of this invention is as follows:

[0127] The system collects and dynamically adjusts environmental and operational parameters; this is achieved through the data acquisition module in this invention. Here, dynamic adjustment refers to adjusting the aforementioned sampling parameters based on the environmental parameters.

[0128] Parameter adjustment and data fusion correction are achieved through the data acquisition module in this invention. The parameter adjustment and data fusion correction mentioned above are the actions of correcting and repairing abnormal running data and setting corresponding weights for running parameters.

[0129] The lifespan loss of components is calculated; this is achieved through the data analysis module in this invention, which executes the lifespan loss calculation model.

[0130] The remaining lifespan of the device to be predicted is obtained by subtracting the life loss from the ideal lifespan of the device to be predicted; this is achieved by the lifespan prediction module in this invention, which is based on the calculation model of the remaining lifespan described above.

[0131] The computational model is trained using batch gradient descent; this is implemented through the adaptive learning module in this invention and used to update the computational model.

[0132] The importance of features is evaluated using the concept of random forests; this is implemented through the adaptive learning module in this invention, which is used to set the relevant weights of the running data.

[0133] The model parameters are updated by adjusting the results, and training is repeated until the model parameters no longer change; this is achieved through the adaptive learning module in this invention, which is the process flow for updating the computational model.

[0134] The present invention also proposes an air conditioning system having the above-mentioned life prediction system.

[0135] Compared with the prior art, the present invention has at least the following beneficial effects:

[0136] The life prediction system proposed in this invention has a data acquisition module, a data analysis module, and a life prediction module. It can collect environmental parameters and operating parameters of the device to be predicted, calculate the life loss of the device to be predicted, and finally estimate the remaining life of the device to be predicted. When this invention is applied to an air conditioning system, it can be used to calculate the life loss of each component in the air conditioning system and predict the remaining life of each component accordingly, so as to carry out maintenance and replacement in advance and effectively extend the service life of the air conditioning system.

[0137] Furthermore, the life prediction system proposed in this invention also has an adaptive learning module, which can optimize the calculation models in the data analysis module and the life prediction module based on a large amount of operational data, thereby creating a calculation model that can accurately predict the remaining life of the device to be predicted, thus improving the accuracy of the prediction of the remaining life of the device to be predicted in this invention.

[0138] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A lifespan prediction system, characterized in that, include: The data acquisition module is used to collect the operating data of the device to be predicted; The data analysis module is used to calculate the life loss of the device to be predicted based on the operating data. A lifespan prediction module is used to estimate the remaining lifespan of the device to be predicted based on the lifespan loss. The data acquisition module includes a first acquisition device for acquiring environmental parameters of the device to be predicted, and a second acquisition device for acquiring operating parameters of the device to be predicted. The sampling parameters of the second acquisition device are set according to the environmental parameters; When acquiring the running data, the data acquisition module performs the following actions: The weights of the operating parameters are set according to the environmental parameters; When acquiring the running data, the data acquisition module also performs the following actions: Determine whether the currently collected running data exceeds the threshold range; wherein, the threshold range is set by 15% deviation of the previous collected data and 15% deviation of the subsequent collected data; If so, then the currently collected running data will be corrected and repaired; The sampling parameters include at least sampling time, sampling frequency, and anti-interference capability; The sampling parameters of the second acquisition device are set according to the environmental parameters, specifically including: dynamically adjusting the sampling frequency according to changes in the environmental parameters, and increasing the sampling frequency when the environment changes drastically in order to capture detailed information; The lifetime prediction system also includes an adaptive learning module, which is used to train and optimize the calculation model for calculating lifetime loss and the calculation model for calculating remaining lifetime using the batch gradient descent method. The adaptive learning module uses the idea of ​​random forest to evaluate feature importance, standardizes the importance of each feature, and sums them to 1 to dynamically adjust the weights of the running parameters. The adaptive learning module updates the model parameters based on the adjustment results and repeats the training until the model parameters no longer change. The training model for the batch gradient descent method is as follows: ； in, , For a single selected sample, Let be the gradient of the loss function for a single sample, and m be the number of samples used for training. Let be the hypothesis function in the batch gradient descent method. This is the weight vector.

2. The lifespan prediction system according to claim 1, characterized in that, The calculation model used by the data analysis module to calculate the lifespan loss of the device to be predicted is as follows: y=β0+β1x1+β2x2+……+βpxp+ε Where y is the lifespan loss, x1, x2, ..., xp are the operating data, β0, β1, ..., βp are the proportional coefficients, and ε is the error value.

3. The lifetime prediction system according to claim 1, characterized in that, The calculation model used by the life prediction module to estimate the remaining life of the device to be predicted is as follows: F=Ny; Where F is the remaining lifetime, N is the ideal lifetime of the device to be predicted, and y is the lifetime loss.

4. An air conditioning system, characterized in that, The air conditioning system has a life prediction system as described in any one of claims 1 to 3.

Images