Method and system for evaluating field device data reliability

Abstract

The invention relates to a method and a system for evaluating field device data reliability. The invention relates to a computer-implemented method for evaluating the reliability of field device data, comprising the steps of: starting (100) a separate cycle; receiving (102) reliability-related data from the field device; calculating (104) and accumulating a first reliability value and / or a first hazard value without compression over an undetermined time span according to a model using at least one variable representative of the reliability-related data; compressively calculating (106) and accumulating a second reliability value and / or a second hazard value over an undetermined time span according to a model using at least one variable representative of the reliability-related data, the value of the at least one variable being compressed; the first and second reliability values and / or second hazard values are compared (108) with associated predetermined thresholds corresponding to accuracy.

Description

Technical Field

[0001] The present invention relates to a method and system for evaluating the reliability of field equipment data under measured variable conditions. Background Technology

[0002] This invention relates to the fields of reliability analysis, predictive maintenance, or more generally, predictive and health management, and asset management systems. It particularly relates to equipment deployed in various field conditions, with power distribution components or systems or electronic devices representing the primary focus.

[0003] In many of these applications, the high reliability required presents challenges in assessing the reliability of the devices in question. This makes laboratory testing challenging, and the dependence on environmental variables increases the amount of testing required. One approach discussed in the literature is to use data from field devices. This approach addresses the aforementioned data scarcity due to the large number of devices in use and the diverse environmental conditions they are exposed to. However, a challenge is that conditions are often variable, necessitating the storage of data from a large number of devices over extended periods. As an illustrative example: measurements taken every minute over a 10-year period yield approximately 5 million data points. This number must then be multiplied by the number of controlled devices, which could be, for example, 10,000 devices. Therefore, the data must be reduced. However, this reduction must not result in unacceptable inaccuracies in the reliability assessment. Summary of the Invention

[0004] It is desirable to provide an improved method for reducing reliability-related data from field devices, such as environmental or operating condition data, in order to assess the reliability of field devices.

[0005] This problem is addressed by the subject matter of the independent claims. Examples are provided by the dependent claims, the following description, and the accompanying drawings.

[0006] The described embodiments similarly relate to methods and systems for evaluating the reliability of field device data, as well as computer program elements and computer-readable media. Synergistic effects can arise from different combinations of implementation schemes, although they may not be described in detail.

[0007] Furthermore, it should be noted that all embodiments of the method according to the present invention can be performed in the order of the described steps. However, this is not necessarily the only and fundamental order of the steps of the method. Unless otherwise expressly stated below, the method presented herein can be performed in a different order of the disclosed steps without departing from the corresponding method embodiment.

[0008] Technical terms are used according to their common sense. If a particular meaning is conveyed to certain terms, the definition of the term will be given in the context of its use below.

[0009] In a first aspect, a computer-implemented method for evaluating the reliability of field equipment is provided. The method includes the following steps: In a first step, a separate loop is initiated. In a second step of the separate loop, reliability-related data for at least one usage and / or environmental variable is received from the field equipment. In a third step of the separate loop, a first reliability value and / or a first hazard value are calculated and accumulated uncompressed over an undetermined time span. The calculation and accumulation are performed according to a model using at least one variable representing the reliability-related data. In a fourth step of the separate loop, a second reliability value and / or a second hazard value are calculated and accumulated compressed over an undetermined time span, according to the model using at least one variable representing the reliability-related data, wherein the value of at least one variable is compressed. In a fifth step of the first loop, the first reliability value and the second reliability value and / or the second hazard value are compared with a predetermined threshold corresponding to accuracy. In a sixth step of the first loop, if the threshold is exceeded, the calculation of the first and second reliability values ​​and / or hazard values ​​is stopped, and the time span is determined. In a seventh step of the separate loop, preferably along with a timestamp, the compressed value of at least one variable is stored, and the separate loop is closed, i.e., the separate loop ends.

[0010] In other words, a reliability value and / or hazard value are calculated once using compressed data and once using uncompressed data. The resulting values ​​are compared to each other. If their difference is too high but still within predetermined limits, the compressed value for the current cycle is stored, and value compression is stopped. This method guarantees the accuracy of the data and the reliability of the device when using compressed data. This method also allows storing compressed values ​​instead of every single value received. Note that the term compressed value is used in this disclosure, where the value is represented by data. Therefore, the expressions "compressed data" and "compressed value" have the same meaning herein.

[0011] Data reduction is addressed through data compression. However, this method provides dynamic compression, rather than compression based on fixed time intervals. This can be inefficient if the intervals are too short, as they may not reflect the shortest conceivable variation in conditions, and potentially introduce unknown errors if the intervals are too long. In other words, data is not aggregated at fixed time intervals (e.g., daily, weekly, or yearly), but rather based on changes in environment and usage conditions.

[0012] Therefore, a method is provided for compressing and storing reliability-related data from a large number of devices in the field. This method allows for the reanalysis of the reliability of these devices in their individual use and environment when the model parameters of the underlying life model change, or alternatively, when combined with field failure or repair data to improve the accuracy of these model parameters. This is important because storing all individual measurements is impractical due to their individual size and the large number of devices. To be most efficient in such storage, a dynamic compression scheme is proposed, which controls the accuracy of the compressed data and immediately saves it once there is a significant change compared to the complete, i.e., uncompressed data. The method assesses the error introduced by compression and then cuts off or restarts compression if a certain error threshold is reached. Compressed data also helps to simplify the reanalysis of devices or models.

[0013] According to the embodiments, reliability-related data are usage and / or environmental data, and the variables representing reliability-related data are the corresponding usage and / or environmental variables.

[0014] The reliability-related data that is compressed and stored can come from device usage and environmental data.

[0015] According to an embodiment, at least one variable is related to temperature, pressure, humidity, voltage, current, pH value, vibration, the intensity of salt or corrosive substances, or the amount of chemical substances.

[0016] In other words, environmental and usage data can contain one or more of these data types. Each data type is assigned to a corresponding variable, which accepts the corresponding data as its value. This list is not exhaustive.

[0017] According to an embodiment, the model includes an acceleration factor, which depends on at least one usage and environment variable as well as model parameters.

[0018] In the context of accelerated life modeling, an acceleration factor describes the increase or decrease in failure rate or time to failure under specific conditions that differ from some reference conditions. The time to failure of equipment is typically related to the amount of stress applied to the equipment. Acceleration factors are used to mathematically model this correlation among various failure mechanisms. Typically, the reference failure rate or time to failure is determined by testing under reference conditions, and an increase in the failure rate or a decrease in the time to failure is determined by testing under specific conditions. By using acceleration factors, higher stresses can be applied to the equipment in normally performed tests than in the field, and the time to failure or cycle in the test can be shortened. In this disclosure, field data itself is used instead of testing under specific conditions. This is possible because, among a large number of devices, many are exposed to environmental and operational conditions that correspond to higher stress conditions in the test.

[0019] Model parameters can be adapted during device operation. This can occur, for example, every few years, and independently of the cyclic and reliability assessments presented in this disclosure. Updates to model parameters can be based on additional information, such as observed failures and their causes. However, updates can be synchronized with the end or start of a cycle. Therefore, as an option, a separate cycle may include a step to check if model parameters need updating, and if so, to combine data from a large number of devices, including failure or service report data, and reanalyze the model parameters. The model parameters can then be adapted. Furthermore, the updated model parameters can be used to reassess reliability. By adapting the model parameters, the model is improved and refined. The updated model is used to perform calculations and comparisons of reliability values ​​and / or hazard values ​​using compressed and uncompressed data. The new model parameters can also be validated by using them not only for the next cycle but also for the previous cycle. For this purpose, stored compressed values ​​can be used.

[0020] To control accuracy, thresholds can be predetermined, representing allowed values ​​or ranges of values ​​that model parameters can or may adopt in future adaptations. That is, not only is accuracy proven, but as a secondary criterion, the range of parameters is independent of accuracy. For example, one or more representative constraints on the parameters in the model are predetermined. These are typically derived from prior knowledge or known physical constraints.

[0021] According to an embodiment, the compressed data, i.e., the compressed values, are used to calculate reliability, which is a function of the time covered or alternatively as a specific time instance.

[0022] According to an embodiment, when the parameters of the model have changed, the compressed and stored values ​​are used to reassess reliability at a fixed time or as a function of time.

[0023] According to an embodiment, a further step includes receiving isolated fault or repair information from the field device. The compressed value, along with the isolated fault or repair information, is used to update or improve model parameters.

[0024] For example, separate fault or repair information can be applied when or before a separate loop closes. The updated and improved model parameters for the next loop can be based on all available compressed values ​​along with the fault and repair information, where "all available compressed values" refers to all values ​​stored over the lifetime of a single device. Therefore, it includes all previous compressed values ​​and can be part or all of the current loop's values. For example, maximum likelihood estimation (MLE) can be used to update the model parameters to maximize the model's likelihood function.

[0025] According to an embodiment, compression is performed using one or more statistical methods, accuracy reduction, truncation, or downsampling.

[0026] Compression can be performed using any known mathematical algorithm or technique suitable for the intended application. In particular, statistical methods, downsampling (i.e., omitting values ​​according to rules or statistics), reduced accuracy, or truncation can be used, which relates to the reduction of significant digits in the data representing the values.

[0027] According to an embodiment, one or more statistical methods are used to estimate one or more key statistical properties, particularly the moments of a distribution, such as the mean or mean, standard deviation or variance, or potentially higher-order moments.

[0028] By statistically distributing values ​​using assumed parameters, the parameters of the distribution can typically be used. That is, instead of using statistical results such as the mean, the values ​​of these parameters represent compressed values. This list is not exhaustive.

[0029] As an example, a simple method can be used to calculate the average temperature over a time interval. Other times and periods are also suitable. Alternatively, the mean and standard deviation can be calculated based on possible values ​​or a range of variables (e.g., all true values, only positive values, or values ​​in the interval, such as [0,1]) and used to fit a suitable statistical distribution (e.g., normal, gamma distribution, beta distribution). The cumulative hazard function or change in reliability is then based on the use of that distribution.

[0030] According to an embodiment, data from field devices is collected over a complete cycle, and the collected data over the complete cycle is used to perform compression.

[0031] A complete cycle comprises the complete, individual cycles as described above. As mentioned, a complete cycle begins at the end of the previous cycle and ends at the end of the current cycle. Therefore, a complete cycle includes the defined time span.

[0032] According to an embodiment, data from the field device is continuously compressed during the complete cycle.

[0033] For example, a period is defined on which statistical values ​​(e.g., averages) are calculated and used as compressed values. Since the time intervals can become longer, this embodiment allows for a reduction in computational workload at the end of the cycle.

[0034] According to the embodiments, the accuracy standard is a relative value or an absolute value.

[0035] In other words, the reliability and / or hazard values ​​are compared to the threshold by dividing the reliability and / or hazard values ​​of the compressed and uncompressed data and comparing the result with the threshold, or by subtracting them and comparing the result with the threshold.

[0036] According to one aspect, a system for evaluating the reliability of field device data is provided. The system includes a computing device connected to a plurality of field devices, wherein the computing device is configured to perform the steps of the method according to any of the foregoing claims.

[0037] A computing device includes at least one processor, memory for storing a program, and memory for storing data, the program containing instructions for performing the method. The computing device can be a single hardware device, a distributed device, or a virtual device.

[0038] The system may also include multiple field devices. The connection between the computing device and the field device can be achieved in a wired or wireless manner through one or more of the following physical hardware and hardware and / or software protocols: fieldbus, Ethernet, Bluetooth, wireless LAN, LTE, GSM, to name a few, or any wireless protocol and / or any wired protocol, and the corresponding communication unit.

[0039] A computer program element can be a part of a computer program, but it can also be the entire program itself. For example, a computer program element can be used to update an existing computer program to obtain the present invention.

[0040] According to another aspect, program elements are provided, which include instructions for performing steps of the method.

[0041] According to another aspect, a computer-readable storage device on which program elements are stored.

[0042] Computer-readable media can be considered as storage media, such as, for example, USB sticks, CDs, DVDs, data storage devices, hard disks, or any other media on which program elements as described above can be stored.

[0043] Although the current description deals with reliability as a function of time, it is well understood that the use of time can be replaced by other variables measuring the aging process of the equipment. Commonly used alternatives are especially the number of operations or loops performed when the system operates irregularly over time.

[0044] By studying the accompanying drawings, the disclosure, and the appended claims, those skilled in the art who practice the claimed invention can understand and implement other variations of the disclosed embodiments. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude multiple. A single processor or other unit can perform the functions of several items or steps listed in the claims. The fact that certain measures are listed in mutually different dependent claims does not indicate that a combination of these measures cannot be used advantageously. Computer programs can be stored / distributed on suitable media, such as optical storage media or solid-state media provided with or as part of other hardware, but can also be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems. Any reference numerals in the claims should not be construed as limiting the scope of the claims.

[0045] These and other features, aspects and advantages of the invention will be better understood with reference to the accompanying drawings and the following description. Attached Figure Description

[0046] Figure 1 A flowchart is shown of a computer-implemented method for evaluating the reliability of field device data.

[0047] Figure 2a The first scenario is shown, with different conditions represented by the environmental variable x.

[0048] Figure 2b The results of the compressed and uncompressed methods for the first scenario with respect to the effective time t_effective are shown.

[0049] Figure 3a A second scenario is shown with different conditions represented by the environmental variable x.

[0050] Figure 3b The results of the compressed and uncompressed methods for the second scenario with respect to the effective time t_effective are shown.

[0051] Figure 4 A block diagram of the system is shown. Detailed Implementation

[0052] In all the accompanying drawings, corresponding parts have the same reference numerals.

[0053] Figure 1 A flowchart of a computer-implemented method 100 for evaluating the reliability of field device data is shown.

[0054] The complete loop consists of step 102, where the loop begins, and step 116, where the loop ends. Within the loop, steps 102 to 110 repeat until the condition described below is met. When the condition is met, steps 112 to 116 are executed, ending the current loop and starting a new loop.

[0055] Starting from the first cycle, reliability-related data from the field equipment is received at step 102. This reliability data can be environmental data and / or usage data related to the equipment environment and / or equipment usage. The data is associated with at least one variable corresponding to the environmental type, such as temperature, humidity, pressure, pH, etc. Data related to the environmental type is also referred to herein as "data type".

[0056] In step 104, a first reliability value and / or a first hazard value are calculated and accumulated over an undetermined time span. The data used in this step is uncompressed data. The calculation and accumulation are performed according to a model using at least one variable representing the reliability-related data. An example is also given below. In the third step of the first loop, a second reliability value and / or a second hazard value are calculated and accumulated compressively over an undetermined time span according to a model using at least one variable representing the reliability-related data, wherein the value of at least one variable is compressed.

[0057] In another step 106, which may be performed before, after, or concurrently with step 104, compressed data is used to calculate and accumulate a second reliability value and / or a second hazard value over an undetermined time span. The calculation and accumulation are performed according to the model.

[0058] In step 108, the first reliability value and the second reliability value and / or the second hazard value are compared with a predetermined threshold corresponding to accuracy. To do this, a ratio or difference between these values ​​is calculated such that the comparison is relative or absolute.

[0059] In step 110, a decision is made as to whether to repeat steps 102 through 108. This is the case if the comparison result of the previous step did not exceed the threshold. Otherwise, if the threshold has been exceeded, the calculation of the first reliability value, the second reliability value, and / or the hazard value is stopped, and the time span is determined. That is, the event defines the end time of the first cycle. In the next step 112, the compressed value of at least one variable is stored, and in step 114, the first cycle is closed. Simultaneously, the next cycle begins, and execution jumps back to step 102.

[0060] Suppose that the equipment in the field measures at least one variable that affects its reliability (e.g., temperature, humidity, voltage, current level, etc.). This data relating to one of these measured variables is also referred to in this disclosure as environmental type data or a "data type". Hereinafter, we consider a single variable. Extensions to more than one variable, or even variables that interfere with each other, are straightforward.

[0061] This variable is used to relate it to the acceleration factor using an acceleration function. The acceleration function depends on more variables and parameters as described below. The impact of these "acceleration factors" (AF) on the hazard rate, or alternatively, reliability, is described using common models. For example, two of the most commonly used models are the "Scale Accelerated Time to Failure" (SAFT) model, also known as the "Accelerated Lifetime" (ALT) model, and the "Proportional Hazard" (PH) model. Their formulas for effective lifetime or modified hazard functions are given below. For example, effective lifetime is used as a model, which is given as...

[0062]

[0063] Where x(t) represents the environmental impact, the accelerated function form of AF(x; θ) depends on x and some model parameters θ. Alternatively, the cumulative hazard function can be used, which is given by...

[0064]

[0065] Where h′0(t) is the basic or reference hazard function, and as before, AF(x;θ′)AF′(x;θ′) is the acceleration factor.

[0066] One challenge in the practical application of this model is determining the necessity of the acceleration factor AF. While the functional form of AF is known in many cases—for example, using Arrhenius, Peck-Hallberg, or inverse power-law forms—the parameters in these models must be determined. This is typically done in specialized laboratory experiments, which is both time-consuming and expensive. An alternative approach is to utilize field data, where environmental variables or usage are measured and subsequently combined with, for example, fault information from reports, to evaluate them. For highly reliable equipment, and for large numbers of devices, data must be collected and stored over extended periods. This results in a large volume of data, posing a challenge for practical applications. One solution is to compress the data. This can be achieved, for example, by reducing or averaging it over a longer time span. Alternatively, statistical properties such as mean and standard deviation can be calculated to capture its variability. However, conditions are often not stable. Therefore, the compression algorithm must be informed of the time span over which compression should be applied. To address this issue, the method proposed in this paper uses one or more models describing lifetime acceleration. Furthermore, a representative value or two or more limits for the parameters in the model are determined. These can be derived from prior knowledge or known physical constraints. Using uncompressed data, the cumulative harm or relative change in reliability is calculated based on the compressed data with parameter values. Compression is stopped based on a predetermined accuracy standard, and the current compressed value and time interval are sent or stored. Then, a new compression cycle begins.

[0067] For compression algorithms, many methods can be employed: a simple approach is to use the average of the variable over a time interval alone. Other times are also suitable. Alternatively, based on the range of possible parameters of x, the mean and standard deviation are calculated and used to fit a suitable statistical distribution (e.g., normal, gamma, beta distribution). Such a range can include, for example, all possible values, only positive values, or be limited to an interval, such as [0,1]). The cumulative hazard function or change in reliability is then based on the use of this distribution.

[0068] As a concrete example, we assume, for instance, that the mean μ(t) and standard deviation σ(t) are calculated over time intervals [t, t+t]. The use of compressed data for, for example, assessments of effective age includes...

[0069]

[0070] in It is an approximation based on the mean and standard deviation, x(t′), t′∈[t,t+T], and θ is a typical value or one of the parameter values ​​used for evaluation.

[0071] As a simple example, assume the compression involves using only the average value μ(x) of the data. Calculation Including direct use Instead of uncompressed values.

[0072] An alternative approach is to assume that x follows a normal distribution and that the acceleration factor has the form exp(-βx) with β as a parameter. Compression involves calculating μ(x) and σ(x). In this case, the acceleration factor is given by the following equation.

[0073]

[0074] Its difference from the pure average method lies in the second term in the exponent.

[0075] For accuracy standards, absolute or relative standards can be used. Additionally, a minimum time interval can be used (e.g., to capture at least daily changes). In any case, the maximum time interval after which data can be stored can also be used.

[0076] If the parameters are known more accurately, the stored data can then be used to reassess reliability at a later time. Accuracy is guaranteed if the parameters are within the range used for compression. Furthermore, and even more importantly, compressed data can be used in conjunction with separately collected fault or repair information to update or improve model parameters. A typical example is the MLE method, where the likelihood function of the reliability model is maximized.

[0077] Compression algorithms can be implemented in two ways. The first is to store all the information over increasingly longer time intervals. This becomes challenging as the time intervals grow larger. A more suitable approach is an online algorithm, where data is processed simultaneously with measurement.

[0078] Figures 2a to 3b Two examples with simulations are shown, in which the accelerated life model is used as the basis along with the power law of the covariate x and the value of the power as an unknown parameter.

[0079] Figure 2a and Figure 2b An example with a first scenario is shown, where the conditions change at one moment, and Figure 3a and Figure 3b A second example is shown, where conditions change gradually over time. The mean and standard deviation are used for compression. Figure 2b and Figure 3b The table shows the results of correctly calculated validity time and validity time using the compressed version.

[0080] Figure 2a The environment variable x is described as a function of time. A sudden change occurs at t = 50. (Example...) Figure 2bAs shown, although the exact calculation (solid line) and the calculation using the compression method (dashed line) are consistent at this point, they subsequently diverge. For example, at t=55, the compression process will be restarted.

[0081] Figure 3a The environment variable x is depicted again as a function of time. This second example shows the gradual change of the environment variable x. As the condition changes more, the difference between the exact (solid line) and the compressed method (dashed line) widens, possibly indicating the need to restore compression.

[0082] Figure 4 A block diagram of a system 400 for evaluating the reliability of field device data is shown. System 400 includes a computing device 402 connected to a plurality of field devices 410, wherein the computing device 402 is configured to perform the steps described herein. The computing device 402 includes at least one processor 404, a first memory 406 for storing a program containing instructions for performing the steps of the method, and a second memory 408 for storing data. The computing device 402 may be a single hardware device or a distributed or virtual device. The connection between the computing device 402 and the field devices 410 can be implemented wired or wirelessly. For this purpose, the computing device includes one or more corresponding communication devices 412. A network with network components such as base stations, routers, gateways, or other network nodes can also be used to implement the connection.

Claims

1. A computer-implemented method for evaluating the reliability of field device data, the method comprising the steps of: Start a separate loop (100); Receive reliability-related data from field equipment (102); Based on a model using at least one variable representing the reliability-related data, calculate (104) and accumulate the first reliability value and / or the first hazard value over an undetermined time span without compression. Based on the model using the at least one variable representing the reliability-related data, a second reliability value and / or a second hazard value are calculated (106) and accumulated over the yet-to-be-determined time span in a compressed manner, wherein the value of the at least one variable is compressed; The first reliability value and the second reliability value and / or the second hazard value are compared with a predetermined threshold corresponding to the accuracy (108); If the threshold (110) is exceeded, then the calculation of the first reliability value and the second reliability value and / or hazard value is stopped (112) and the time span is determined; as well as Storage (114) of the compressed value of the at least one variable; and Close the separate loop described in (116).

2. The method of claim 1, wherein the reliability-related data is usage and / or environmental data, and the at least one variable represents the reliability-related data.

3. The method according to any one of the preceding claims, wherein the at least one variable is related to temperature, pressure, humidity, voltage, current, vibration, amount of chemical substance, intensity of salt or corrosive substance, or pH value.

4. The method according to any one of the preceding claims, wherein the model includes an acceleration factor, wherein the acceleration factor depends on the at least one usage and environmental variable and the model parameters.

5. The method according to any one of the preceding claims, wherein the compressed value is used as a function of the covered time or alternatively for a specific time instance to calculate reliability.

6. The method according to any one of the preceding claims, wherein when the parameters of the model have been changed or alternatively for a specific time instance, the compressed and stored values ​​are used as a function of time to reassess the reliability.

7. The method according to the preceding claim, wherein adapting the model comprises: Separate fault or repair information is received from the field device, and all available compressed values ​​are used together with the received separate fault or repair information to adapt the model parameters.

8. The method according to any one of the preceding claims, wherein the compression is performed using one or more statistical methods, accuracy reduction, truncation, or downsampling.

9. The method of claim 8, wherein the one or more statistical methods are used to estimate parameters of the statistical distribution of the data, or to determine parameters of a hypothetical statistical distribution.

10. The method according to any one of the preceding claims, wherein the data from the field device is collected within a complete cycle, and the compression is performed using the data collected within the complete cycle.

11. The method according to any one of claims 1 to 9, wherein the data from the field device is continuously compressed during the entire cycle.

12. The method according to any one of the preceding claims, wherein the accuracy criterion is a relative value or an absolute value.

13. A system (400) for evaluating the reliability of field device data, comprising a computing device (402) connected to a plurality of field devices (410), wherein the computing device (402) is configured to perform the steps of the method according to any one of the preceding claims.

14. A program element comprising instructions for performing the steps of the method according to any one of claims 1 to 12.

15. A computer-readable storage device having stored thereon the program elements according to claim 14.

Images