A Bayesian-based online fault detection method and system for air conditioner outdoor units
By using a Bayesian data fusion algorithm, combining historical and real-time operating parameters of the air conditioner outdoor unit, the complexity and inaccuracy of existing air conditioner outdoor unit testing methods are solved, enabling fast and accurate online testing that adapts to constantly changing production standards.
Patent Information
- Application Number
- CN202311172967.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-12
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2043-09-12
AI Technical Summary
The existing online testing methods for air conditioner outdoor units have simple criteria and large fluctuations in the standard values, resulting in complex testing methods, inaccurate test results, long testing cycles, and high hardware costs, making it difficult to adapt to constantly changing production standards.
A Bayesian-based multi-sensor online detection method for air conditioner outdoor units is adopted. By acquiring the prior probability of historical operating parameters and fusing it with the posterior probability of real-time operating parameters, a Bayesian data fusion algorithm is used to determine whether the air conditioner outdoor unit is faulty, simplifying the detection steps and improving accuracy.
It has improved the speed, accuracy and efficiency of air conditioner outdoor unit testing, simplified the testing steps, reduced hardware costs, and improved the reliability and adaptability of test results.
Smart Images

Figure CN117308274B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of air conditioner outdoor unit fault detection technology, and more specifically, relates to an online fault detection method and system for air conditioner outdoor units based on Bayesian methods. Background Technology
[0002] Air conditioning product testing is typically divided into factory inspection, sampling inspection, and type inspection. Factory inspection of refrigeration equipment refers to the requirement that each product be inspected individually before leaving the factory. However, some refrigeration equipment products, due to their outdated design, are unsuitable for current production models, making factory inspection difficult to implement or impacting production line rhythm. Existing online factory inspection methods for air conditioners rely on independent comparisons of power and refrigerant pressure deviations to determine if the outdoor unit is faulty. This method is simplistic and unreasonable, as the testing standard for each outdoor unit is based on the average power data of dozens of previously measured air conditioners. However, this average value fluctuates significantly depending on the number of units tested. Furthermore, online testing requires multiple heating and cooling cycles for the outdoor unit, resulting in a long testing cycle and cumbersome procedures. Due to the long testing cycle, a large number of workstations need to be configured in parallel to match the assembly line rhythm, leading to significant space requirements, high hardware costs, and complex software maintenance.
[0003] Furthermore, with the continuous development of advanced testing methods, testing standards are also constantly changing. There is an urgent need for a method that can adapt to the current production methods and standards for air conditioner outdoor units, in order to cope with the ever-changing air conditioner production and testing standards. Summary of the Invention
[0004] To address the shortcomings of existing technologies, the present invention aims to provide a Bayesian-based online fault detection method and system for air conditioner outdoor units, thereby solving the problems of simple judgment criteria, large fluctuation range of judgment standard values, and inaccurate detection results in existing online detection methods for air conditioner outdoor units.
[0005] To achieve the above objectives, according to one aspect of the present invention, a Bayesian-based multi-sensor online detection method for air conditioner outdoor units is provided, comprising the following steps:
[0006] S1 obtains the probability distribution of various evaluation features corresponding to each historical operating parameter of the outdoor unit of the air conditioner under test, as a prior probability.
[0007] S2 calculates the probability of each evaluation feature corresponding to each real-time operating parameter based on each real-time operating parameter acquired multiple times during a first preset time period of stable operation of the outdoor unit of the air conditioner under test, and then performs Bayesian data fusion with the prior probability to obtain the posterior probability.
[0008] S3 compares the posterior probability with a preset threshold, and determines whether the outdoor unit of the air conditioner under test is faulty based on the comparison result.
[0009] Furthermore, the evaluation features include at least qualified features and unqualified features.
[0010] Furthermore, in step S1, if the probabilities of the qualified and unqualified characteristics of the historical operating parameters are the same, then the evaluation characteristics of the historical operating parameters are classified as unqualified.
[0011] Furthermore, in step S3, if the posterior probability is greater than the preset threshold, then the operating parameters corresponding to the outdoor unit of the air conditioner under test are deemed qualified; otherwise, they are deemed unqualified.
[0012] Furthermore, in step S3, the result of the qualification or non-qualification judgment is added to the historical data to update the prior probability.
[0013] Furthermore, in step S1, the mathematical model for statistical analysis of various operating parameters of the outdoor unit of the air conditioner under test within a pre-collected continuous time period using the Bayesian algorithm is as follows:
[0014]
[0015] In the formula, Y represents the category of the sample, K is the number of evaluation feature categories for the outdoor unit of the air conditioner, and c k This indicates the category of the output y. For example, when K=2, k=1,2; i represents the i-th dataset, and N is the total number of input datasets. λ is an indicator function, indicating that the result is 1 when the condition in parentheses is met, and 0 otherwise; λ≥0, which is equivalent to assigning a positive number to the statistical frequency.
[0016] Furthermore, in step S2, the mathematical model for calculating the probability of various evaluation features is as follows:
[0017]
[0018] Where j is the number of sensors, and j = 1, 2, 3...n; S j It is the set of values that each element of the sample point can take. The number of; l represents dividing the continuous data collected by different sensors into l categories.
[0019] Furthermore, in step S3, the mathematical model for calculating the posterior probability is as follows:
[0020]
[0021] Where Y represents the category of the operating parameter sample; K is the number of evaluation feature categories for the outdoor unit of the air conditioner; and c kThe output y represents the category, such as k=1,2 when K=2; N is the number of types of input parameters; R is the number of types of parameters; X represents an n-dimensional vector, R is the number of types of parameters, and x represents the given input.
[0022] Furthermore, the operating parameters include at least pressure and current; preferably, in step S2, the method for obtaining multiple real-time operating parameters within a preset time period during which the outdoor unit of the air conditioner under test is stably operating is as follows: after the outdoor unit of the air conditioner under test has been stably operating for a first preset time period, multiple real-time operating parameters are obtained second by second within a second preset time period; more preferably, the second preset time period has the same length as the first preset time period.
[0023] According to another aspect of the present invention, an online fault detection system for an air conditioner outdoor unit based on Bayesian methods is also disclosed, comprising:
[0024] The prior probability acquisition module obtains the probability distribution of various evaluation features corresponding to each historical operating parameter of the outdoor unit of the air conditioner under test, based on each historical operating parameter of the air conditioner under test, as a prior probability.
[0025] The posterior probability acquisition module calculates the probability of various evaluation features corresponding to each real-time operating parameter based on each real-time operating parameter acquired multiple times during the first preset time period of stable operation of the outdoor unit of the air conditioner under test. Then, it performs Bayesian data fusion with the prior probability to obtain the posterior probability.
[0026] The judgment module compares the posterior probability with a preset threshold and determines whether the outdoor unit of the air conditioner under test is faulty based on the comparison result.
[0027] Compared with the prior art, the above technical solutions conceived by this invention have the following main advantages:
[0028] 1. This invention rapidly and accurately analyzes and processes various historical operating parameters of air conditioner outdoor units to obtain prior probabilities corresponding to various evaluation features of historical operating parameters. Then, under the factory testing environment of the air conditioner outdoor unit assembly line, after the air conditioner outdoor unit under test has been running steadily for a preset time period, real-time operating parameters of each type of air conditioner outdoor unit under test are collected multiple times during a certain period of subsequent operation to obtain complex heterogeneous data. The frequency distribution of different evaluation features corresponding to a series of collected heterogeneous data is then calculated. The calculated frequency distribution and prior probabilities are then fused with Bayesian data to obtain posterior probabilities. By distinguishing different types of outdoor units and defining the range of prior probability judgment, the complexity of conventional data fusion is overcome, making the air conditioner outdoor unit testing steps simpler, reducing testing time, and further improving the efficiency and accuracy of online testing of air conditioner outdoor units.
[0029] 2. This invention provides a mathematical model and a calculation method for calculating frequency distribution, conditional probability, and posterior probability during online testing of air conditioner outdoor units. These calculation methods introduce process detection values that change over time. The detected process detection values (which are within a second preset time period or a specific duration, typically such as the next 30 seconds after the air conditioner outdoor unit has been running stably for 30 seconds, with one detection value taken per second for each operating parameter) are used to calculate the operating status of the air conditioner outdoor unit. This allows for a more accurate judgment of whether the air conditioner outdoor unit is qualified or not. Furthermore, the testing method is simple, fast, and has higher testing efficiency.
[0030] 3. This invention monitors and adjusts the operating status of the outdoor unit of the air conditioner in real time, collects real-time operating parameter data, performs calculations, and saves the calculation data to a local database. That is, when the outdoor unit of the air conditioner is tested, the Bayesian prior probability is updated by recording the calculation data of each newly tested outdoor unit. This can improve the safety and stability of the outdoor unit of the air conditioner and reduce the probability of the outdoor unit of the air conditioner failing again after leaving the factory. Attached Figure Description
[0031] Figure 1 A schematic diagram of the Bayesian-based online fault detection method for air conditioner outdoor units provided by the present invention;
[0032] Figure 2 This is a schematic diagram of the calculation process in the Bayesian-based online fault detection method for air conditioner outdoor units provided by the present invention;
[0033] Figure 3 This is a schematic diagram of the online fault detection process of the air conditioner outdoor unit in Embodiment 1 of the present invention;
[0034] Figure 4 This is a schematic diagram of the fault judgment result of the air conditioner outdoor unit based on the Bayesian fusion recognition algorithm in Embodiment 1 of the present invention;
[0035] Figure 5 This is a schematic diagram of the hardware structure for detecting multiple detection stations of an air conditioner outdoor unit in an embodiment of the present invention.
[0036] Figure 6 This is a schematic diagram of the hardware structure of a single detection station for an air conditioner outdoor unit in an embodiment of the present invention;
[0037] Figure 7 This is a schematic diagram of the internal operation process of the air conditioner outdoor unit detection system in an embodiment of the present invention. Detailed Implementation
[0038] To make the objectives, technical solutions, and advantages of this invention clearer, the 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 and not intended to limit the invention.
[0039] like Figure 1 As shown, this embodiment of the invention provides a Bayesian-based online detection method for multiple sensors on an air conditioner outdoor unit, comprising the following steps:
[0040] Based on historical data, S1 uses the Bayesian algorithm to statistically analyze various operating parameters of the outdoor unit of the air conditioner under test within a pre-collected continuous time period to obtain the probability distribution of multiple evaluation features required for each operating parameter, and uses the probability distribution as the prior probability.
[0041] Then determine whether the running time of the outdoor unit of the air conditioner under test meets the preset time. If it does, proceed to the next step; otherwise, end the test.
[0042] S2 acquires multiple real-time operating parameters multiple times during the first preset time period of stable operation of the outdoor unit of the air conditioner under test, calculates the probability of various evaluation features corresponding to each real-time operating parameter, and then performs Bayesian data fusion with the prior probability to obtain the posterior probability; the typical value of the aforementioned first preset time period is 30s.
[0043] In a preferred embodiment, the method for acquiring multiple real-time operating parameters during a first preset time period of stable operation of the outdoor unit of the air conditioner under test is as follows: after the outdoor unit of the air conditioner under test has been operating stably for a first preset time period, multiple real-time operating parameters are acquired second by second during a second preset time period; in a more preferred embodiment, the aforementioned second preset time period has the same duration as the aforementioned first preset time period, such as 30s, 35s, 38s, etc.
[0044] S3 compares the posterior probability with the corresponding preset threshold to determine whether the outdoor unit of the air conditioner under test is faulty; specifically, if the aforementioned evaluation features are only of two types, namely qualified and unqualified, if the judgment result is qualified, the outdoor unit of the air conditioner is not faulty, and if the judgment result is unqualified, the outdoor unit of the air conditioner is faulty.
[0045] Specifically, the detection principle of the above detection method is as follows: Assuming that in a random inspection of an air conditioner outdoor unit, one and only one of n mutually exclusive events A (i.e., multiple evaluation characteristics, such as pass or fail) must occur, then:
[0046]
[0047] Among them, A i Indicates the i-th event;
[0048] Let B (the various characteristic physical quantities of the outdoor unit of the air conditioner, i.e., multiple operating parameters) be any event other than the above n events. Then, according to the definition of conditional probability and the law of total probability, we have:
[0049]
[0050] p(A1), p(A2), ..., p(A n ) represents A1, A2, ..., A n The probability of an event occurring is a fact known before detection; this knowledge is called "prior information" (existing detection data). In this embodiment, the prior information has a more definite shape, that is, it is given in the form of a probability distribution, as the prior distribution. In the detection of the air conditioner outdoor unit, event B is observed to have occurred, thus providing a new estimate or understanding of the probability of event B. This knowledge is obtained after detection and is called "posterior knowledge," which is also given in the form of a probability distribution in this embodiment:
[0051] p(A1|B), p(A2|B), ... p(A n |B) (3)
[0052] And mathematical model (3) satisfies the following conditions:
[0053]
[0054] The mathematical model (3) is the "posterior distribution", which integrates prior information and new information provided during the experiment to form the current understanding of the probability. This transformation process from prior information to posterior information is the feature of Bayesian statistics used in this invention.
[0055] In some embodiments, the mathematical model for statistical analysis of various operating parameters of the outdoor unit of the air conditioner under test within a pre-collected continuous time period using the Bayesian algorithm in step S1 above is as follows:
[0056]
[0057] In the formula, Y represents the category of the sample, and c k The output y represents the category (e.g., if k = 1, 2, then c1 represents the category of qualified air conditioners, and c2 represents the category of unqualified air conditioners), i represents the i-th dataset, N is the total number of input datasets, and K is the number of evaluation feature categories for air conditioner outdoor units. λ is an indicator function, indicating that the result is 1 when the condition in parentheses is met, and 0 otherwise; λ≥0, which is equivalent to assigning a positive number to the statistical frequency.
[0058] In some embodiments, the mathematical model for calculating the probability of various evaluation features in step S2 described above is as follows:
[0059]
[0060] Where j is the number of sensors, and j = 1, 2, 3...n; S jIt is the set of values that each element of the sample point can take. The number of; l represents dividing the continuous data collected by different sensors into l categories.
[0061] In some embodiments, the mathematical model for calculating the posterior probability in step S2 is:
[0062]
[0063] Where Y represents the category of the running parameter sample; c k The output y represents the category. For example, when k = 1 or 2, c1 represents the category of qualified air conditioner outdoor unit and c2 represents the category of unqualified air conditioner outdoor unit. N is the number of input operating parameter categories; K is the number of evaluation feature categories of air conditioner outdoor unit; R is the number of operating parameter categories; X represents an n-dimensional vector, R is the number of operating parameter categories, and x represents the given input, which is a multi-dimensional vector composed of the actual acquired real-time operating parameters.
[0064] In some embodiments, the aforementioned evaluation features include at least qualified and unqualified features; in other embodiments, the evaluation features may also include other evaluation categories, making the evaluation results more diverse.
[0065] In some embodiments, in step S1, if the probabilities of the calculated qualified features and unqualified features are the same, then these qualified features are classified into the unqualified feature type; that is, if the probabilities of the qualified features and unqualified features of the historical operating parameters are the same, then all the evaluation features of the historical operating parameters are classified as unqualified, so as to make the evaluation results more accurate.
[0066] In some embodiments, in step S3, if the posterior probability is greater than a preset threshold, the corresponding operating parameters of the outdoor unit of the air conditioner under test are determined to be qualified; otherwise, they are deemed unqualified. The judgment method is accurate and simple.
[0067] In some embodiments, the aforementioned operating parameters include at least two parameters: pressure and current, and may also include operating parameters such as temperature. The more types of operating parameters there are, the more accurate the judgment result will be.
[0068] In some embodiments, in step S3, the result of whether the test is qualified or not is also added to the historical data to update the prior probability, thereby further improving the accuracy of subsequent test results of other air conditioner outdoor units.
[0069] The implementation details of the present invention will be described in detail below through specific embodiments.
[0070] Example 1
[0071] Combination Figure 2-3As shown, in the detection of an air conditioner outdoor unit, it is assumed that there are R information sources (including sensors for voltage, current, pressure, temperature, etc.). These information sources can collect the operating parameter information of the air conditioner outdoor unit under test. The operating parameter information is represented by an R-dimensional vector, and each information source can provide an explanation (i.e., hypothesis) about the target identity based on observation and a specific classification algorithm.
[0072] Information collected from each information source (such as detected voltage, current, temperature, etc.) is divided into K evaluation levels, i.e., K evaluation targets, to represent the hypothesis of the i-th information source about the target's identity. Now, suppose we need to classify the K evaluation levels. The input X is an n-dimensional vector, and the output Y represents the category of the running parameter sample, c. k If we represent the category of output y, then we can obtain:
[0073]
[0074]
[0075] like Figure 2 As shown, the main steps of the Bayesian fusion recognition algorithm include:
[0076] (1) Transform the observations of the target from each of the aforementioned information sources into a hypothesis X about the target's identity;
[0077] (2) Then calculate the uncertainty of the information source regarding the target's identity assumption:
[0078] P(X=x|Y=c k )=P{X (1) =x (1) ,X (2) =x (2) ,…,X (n) =x (n) |Y=c k} (10)
[0079] (3) Finally, calculate the fusion probability of the target identity:
[0080]
[0081] The vectors in X are independent of each other, then
[0082]
[0083] In actual detection, according to the existing historical dataset at the air conditioner outdoor unit detection site, and the evaluation grades of the detection system are determined as two categories, namely qualified and unqualified, that is, the aforementioned K is 2, k=1, 2. In the last few seconds of the stable operation stage of the air conditioner outdoor unit (for example, a typical value of 5 seconds can be taken), one piece of data is acquired per second for each operation parameter, and a total of n pieces of data are acquired. The data results are calculated according to the formulas in the aforementioned steps (1)-(3), that is, the probability distribution of the evaluation grades corresponding to the collected various parameters such as power, pressure and temperature during the operation of the air conditioner outdoor unit is calculated;
[0084] Specifically, in this embodiment, in the multi-sensor online detection of the air conditioner outdoor unit, according to the inherent characteristics of the air conditioner outdoor unit and the on-site detection importance, for the operating state of the air conditioner outdoor unit, three operating parameters of power, refrigerant pressure and temperature are selected as the data collected by each sensor (i.e., information source). When the detection start time reaches the specified time (that is, the operation of the air conditioner outdoor unit in the aforementioned step S2 reaches the preset operation time), the operating state of the outdoor unit tends to be stable, various operating parameters are collected multiple times in a continuous time period, and then according to Figure 2 The calculation process in starts to calculate the probability results of qualification and non-qualification corresponding to the power, pressure and temperature of the air conditioner outdoor unit collected at this time. The calculation result is the posterior probability in the Bayesian model, and then the maximum likelihood estimation method is used to determine whether the outdoor unit is qualified or not and output the classification result (i.e. Figure 2 The statistical decision and fusion identity description in), at the same time, the judgment data of this time is added to the prior probability, and the prior probability is updated in time to improve the accuracy of subsequent air conditioner outdoor unit detection.
[0085] In this embodiment, the algorithm calculation process of the detection method is specifically as follows:
[0086] The Bayesian estimation model of the prior probability P{Y=c k} is constructed as follows:
[0087]
[0088] In the formula, N is the total number of incoming datasets in Figure 3 , the evaluation feature categories of the air conditioner outdoor unit are K=2: qualified and unqualified, k=1, 2; is an indicator function, λ≥0, which is equivalent to assigning a positive number to the statistical frequency. When λ=0, it is equivalent to the maximum likelihood estimation method, and when λ=1, it is Laplace smoothing.
[0089] The Bayesian estimation model of the conditional probability P{X j =x j |Y=c k} is constructed as follows:
[0090]
[0091] Since data from three sensors were collected previously as data sources, the number of data sources j is taken as 1, 2, and 3; where S j It is the set of possible values for each element of the sample point. The number of.
[0092] For a given input x, the posterior probability is calculated using the following formula:
[0093]
[0094] Then output the result of whether it is qualified or unqualified:
[0095]
[0096] If the probabilities of passing and failing are the same, then the passing type will also be classified as the failing type.
[0097] Finally, the output evaluation results are compared with the preset threshold (qualified data of the air conditioner outdoor unit operating parameters determined based on historical data) to determine whether there is a fault or abnormality in the air conditioner outdoor unit, and to handle and repair the air conditioner outdoor unit in a timely manner.
[0098] Specifically, such as Figure 4 The figure shows the output results after testing four different models of air conditioner outdoor units according to the online testing method of this embodiment. The types of operating data collected during the testing process are the power and pressure of each air conditioner outdoor unit. The figure clearly shows the distribution of qualified and unqualified air conditioner outdoor units corresponding to two-dimensional power and pressure. Figure 4 The results for the outdoor unit of the 12800 model air conditioner are in section a. Figure 4 The results for the outdoor unit of air conditioner model 40016 are shown in section b. Figure 4 The results in section c are for the outdoor unit of the 17800 model air conditioner. Figure 4 The image shows the test results for model 12602 air conditioner outdoor unit (d). It is worth noting that the power and pressure parameters of the four types of non-conforming outdoor units (represented by the shaded area indicated by the number x) are located at the boundary. Therefore, these non-conforming units can be easily identified. After testing, special attention and handling are needed for these boundary-located non-conforming units to improve the quality of the air conditioner outdoor units.
[0099] It should be understood that prior to this invention, no testing method for air conditioner outdoor units had adopted the testing method of this invention, nor had any air conditioner outdoor unit testing method or similar equipment factory testing method been able to overcome the problems to be overcome and achieve the effects of this invention. Therefore, this invention, which combines Bayesian data fusion algorithm for factory testing of air conditioner outdoor units, is groundbreaking.
[0100] Example 2
[0101] This embodiment discloses an online fault detection system for air conditioner outdoor units based on Bayesian methods, capable of implementing the detection methods in any of the above embodiments. The system specifically includes:
[0102] The prior probability acquisition module is used to perform statistical analysis on a dataset of multiple operating parameters of the outdoor unit of the air conditioner under test within a continuous time period based on historical data and using the Bayesian algorithm, so as to obtain the probability distribution of multiple evaluation features required for each operating parameter (i.e., the frequency distribution of qualified and unqualified features), and use this probability distribution as the prior probability.
[0103] The posterior probability acquisition module is used to acquire multiple real-time operating parameters multiple times within a preset time period during which the outdoor unit of the air conditioner under test is running stably, calculate the probability of various evaluation features corresponding to each real-time operating parameter, and then perform Bayesian data fusion with the probability of each evaluation feature and the corresponding prior probability to obtain the posterior probability.
[0104] The judgment module is used to compare the preset threshold and the posterior probability to determine whether the outdoor unit of the air conditioner under test is faulty.
[0105] In a preferred embodiment, the aforementioned system further includes a running time determination module, which is used to determine whether the running time of the outdoor unit of the air conditioner under test meets the preset time after the prior probability acquisition module obtains the prior probability. If it meets the preset time, the next step is performed; otherwise, the detection ends.
[0106] The steps for conducting the detection using the above-mentioned detection system include:
[0107] Step 1: Configure the multi-sensor data acquisition device of the air conditioner outdoor unit online monitoring system to collect multiple parameters during the operation of the air conditioner outdoor unit, including pressure, temperature, current and other parameters, and send the collected data to the data processing device for processing.
[0108] Step 2: Statistically analyze the collected data using historical data to obtain the frequency distribution of different parameters, and assign prior probabilities based on the frequency distribution using Bayesian data fusion. These prior probabilities will be used in subsequent Bayesian calculations.
[0109] Step 3: Using the maximum likelihood estimation method, calculate the physical quantities detected by the system, including but not limited to refrigerant pressure and refrigerant temperature, and fuse the calculation results with the assigned prior probability using Bayesian data fusion to obtain the posterior probability. Finally, compare the posterior probability with the preset threshold to output the evaluation result of the air conditioner outdoor unit's quality.
[0110] In other embodiments, the hardware structure of the detection system is as follows: Figure 5 and 6As shown, the detection system is built using PLC modules, KingSCADA software, and Visual Studio development tools. The relevant modules and their inter-module connections are standard techniques used by those skilled in the art and will not be elaborated upon here. Figure 7 The diagram shows the operation flowchart of the detection system in this invention. After the PLC module controls the start of the detection, the system is powered on and then it is determined whether the touch screen display is normal. If it is not normal, the detection system continues to be powered on. If it is normal, the detection data of the air conditioner outdoor unit is checked. Then it is determined whether the detection data of the air conditioner outdoor unit is normal. If it is, the ID code of the air conditioner outdoor unit is scanned to identify the identity of the air conditioner outdoor unit. If not, the power is restarted. After the ID code is successfully scanned, the detection is started. The specific data calculation is performed using the KingSCADA software and the calculation program built into the Visual Studio development tool, and the final data calculation result is displayed on the touch screen on the PLC.
[0111] The above are merely preferred embodiments of the present invention and are 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 Bayesian-based multi-sensor online detection method for air conditioner outdoor units, characterized in that, Includes the following steps: S1 obtains the probability distribution of various evaluation features corresponding to each historical operating parameter of the outdoor unit of the air conditioner under test, as a prior probability; the evaluation features include at least qualified features and unqualified features. If the probabilities of the qualified and unqualified characteristics of the historical operating parameters are the same, then the evaluation characteristics of the historical operating parameters are classified as unqualified. S2 After the outdoor unit of the air conditioner under test has been running stably for a first preset time period, multiple real-time operating parameters are acquired second by second during a second preset time period. Based on each real-time operating parameter acquired multiple times during the first preset time period of stable operation of the outdoor unit of the air conditioner under test, the probability of each type of evaluation feature corresponding to each real-time operating parameter is calculated. Then, the probability is fused with the prior probability using Bayesian data to obtain the posterior probability. S3 compares the posterior probability with a preset threshold, and determines whether the outdoor unit of the air conditioner under test is faulty based on the comparison result; the judgment result of whether it is qualified or not is added to the historical data to update the prior probability.
2. The Bayesian-based online fault detection method for air conditioner outdoor units as described in claim 1, characterized in that, In step S3, if the posterior probability is greater than the preset threshold, the operating parameters of the outdoor unit of the air conditioner under test are deemed to be qualified; otherwise, they are deemed unqualified.
3. The Bayesian-based online fault detection method for air conditioner outdoor units as described in claim 1, characterized in that, In step S1, the mathematical model for statistical analysis of various operating parameters of the outdoor unit of the air conditioner under test within a pre-collected continuous time period using the Bayesian algorithm is as follows: In the formula, Y Indicates the category of the sample. K The number of evaluation feature categories for air conditioner outdoor units. Indicates output y Categories, such as K When =2, k =1,2, i Indicates the first i Data sets, N The total number of input datasets; This is an indicator function, meaning that the result is 1 if the condition in parentheses is met, and 0 otherwise. This is equivalent to assigning a positive number to the statistical frequency.
4. The Bayesian-based online fault detection method for air conditioner outdoor units as described in claim 1, characterized in that, In step S2, the mathematical model for calculating the probability of various evaluation features is as follows: in, j The number of sensors, and j =1, 2, 3... n ; It is the set of values that each element of the sample point can take. The number of; l This represents dividing continuous data collected by different sensors into... l kind.
5. The Bayesian-based online fault detection method for air conditioner outdoor units as described in claim 1, characterized in that, In step S2, the mathematical model for calculating the posterior probability is as follows: in, Y Indicates the category of the running parameter sample; K The number of evaluation feature categories for air conditioner outdoor units. Indicates output y Categories, such as K When =2, k =1,2; N The number of input runtime parameters; R The number of types of running parameters. X Represent a n dimensional vector, R The number of types of running parameters. This indicates the given input.
6. The Bayesian-based online fault detection method for air conditioner outdoor units as described in claim 1, characterized in that, The operating parameters include at least pressure and current, and the second preset time period has the same length as the first preset time period.
7. A Bayesian-based online fault detection system for air conditioner outdoor units, characterized in that, include: The prior probability acquisition module obtains the probability distribution of various evaluation features corresponding to each historical operating parameter of the outdoor unit of the air conditioner under test, based on each historical operating parameter of the air conditioner under test, as a prior probability. If the probabilities of the qualified and unqualified characteristics of a historical operating parameter are the same, then the evaluation characteristic of that historical operating parameter is classified as unqualified. The posterior probability acquisition module acquires each real-time operating parameter second by second during a second preset time period after the outdoor unit of the air conditioner under test has been running stably for a first preset time period. It calculates the probability of each evaluation feature corresponding to each real-time operating parameter, and then performs Bayesian data fusion with the prior probability to obtain the posterior probability. The judgment module compares the posterior probability with a preset threshold, and determines whether the outdoor unit of the air conditioner under test is faulty based on the comparison result. The judgment result of whether it is qualified or not is added to the historical data to update the prior probability.
Citation Information
Patent Citations
Household air conditioner fault diagnosis method based on Bayesian network
CN107806690A
