A method and system for cell partitioning optimization using a general detection tool

Abstract

This invention discloses a unit partitioning optimization method and system using general-purpose detection tools, belonging to the field of electronic equipment fault diagnosis. The invention constructs a candidate partitioning set by traversing the group leader unit and selecting the best member units. The group leader unit is the unit with the longest unit inspection time. When selecting the best member units, several units with the highest probability of failure and no more than the group leader unit's state inspection time are selected from all currently uninspected units. These members, together with the feasible group leader units, constitute a feasible unit partitioning for the j-th detection. The partitioning with the minimum overall equivalent detection time is selected from all candidate feasible unit partitioning methods as the current partitioning, ultimately achieving optimized partitioning of all units. This partitioning optimization method achieves the globally optimal neighborhood solution, significantly reducing the average troubleshooting time.

Description

Technical Field

[0001] This invention belongs to the field of electronic equipment fault diagnosis, and more specifically, relates to a method and system for optimizing unit partitioning using general testing tools. Background Technology

[0002] Electronic device interfaces are easily standardized, and there are testing tools that can detect the same broad category of devices, but with different specific specifications and models; these are called universal testing tools. For example, a device may have multiple hard drives, each loaded with its own programs and data files. When one hard drive develops bad sectors, it can cause intermittent failures. Although the hard drives may not be exactly the same model, they all use standard interfaces. In this case, a universal testing tool can be used to simultaneously check multiple hard drives for bad sectors or perform random data read tests. As equipment / systems become more powerful and advanced, they also become more complex. When a complex equipment / system exhibits a certain malfunction, there are numerous possible causes, making the task of locating the faulty unit extremely demanding.

[0003] Testing tools are an important maintenance resource. The maximum number of units a general-purpose testing tool can test at one time is called its capacity. Generally speaking, the larger the capacity of a testing tool, the faster it can identify the units causing the failure. However, the cost of the testing tool, including its space requirements, will also be greater.

[0004] The decision-making process of selecting a common testing unit for use with a general testing tool when developing a troubleshooting plan is called unit partitioning. A fundamental problem in developing a troubleshooting plan is: how to optimally select multiple test units each time the testing tool is used to quickly locate the faulty unit? This problem has not yet been well resolved in the industry. Especially for complex equipment or systems, the number of test units can range from dozens to over a hundred. Unit partitioning can be viewed as a permutation and combination problem of test units. For example, when the number of units is 16 and the testing tool capacity is 4, the number of partitions exceeds 63 million, making it impossible to effectively optimize the troubleshooting plan using a traversal approach. Currently, unit partitioning mainly relies on the experience of frontline maintenance personnel, resulting in inconsistent partitioning quality and difficulty in providing accurate quantitative results such as the time required to complete troubleshooting, which is detrimental to developing a scientific troubleshooting plan. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to provide a cell partitioning optimization method and system using a universal detection tool, thereby solving the problem that existing cell partitioning methods rely on experience and result in inconsistent partitioning quality.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a method for optimizing cell partitioning using a general-purpose detection tool, the method comprising:

[0007] S1. Obtain the capacity m of the general detection tool, the number of electronic units constituting the electronic device, the time for detecting the status of each electronic unit, and the probability of each electronic unit failing during the task time;

[0008] S2. Round up the ratio of the number of electronic units to the capacity as the maximum number of detections nm, initialize the detection order j = 1, and initialize the set of undetected units as all the electronic units constituting the electronic device;

[0009] S3. Find all feasible group leader units from the current set of undetected units, where the group leader unit is the unit with the longest inspection time; for each feasible group leader unit, select no more than m - 1 units with the highest failure probability and the longest inspection time of the group leader unit status, and jointly form a feasible unit partition for the j-th detection with the feasible group leader unit. All feasible unit partitions jointly form the candidate set of feasible unit partitions for the j-th detection;

[0010] S4. Select the partition method with the smallest overall equivalent detection time from the candidate set of feasible unit partitions for the j-th detection as the unit partition result for the j-th detection, and update the set of undetected units as the difference set between the current set of undetected units and the unit partition result of the j-th detection;

[0011] S5. Update j = j + 1. If j < nm, enter S3; otherwise, use the current set of undetected units as the unit partition result for the last detection.

[0012] Preferably, step S3 includes:

[0013] S31. Initialize the number of groups s = 0, i = 1, and the number of elements na in the current set of undetected units is na = n-(j - 1)*m, where n is the number of electronic units constituting the electronic device;

[0014] S32. Initialize the group leader unit number z = A i , and the remaining elements in the set of undetected units A except z are the candidate group member unit numbers, denoted as At;

[0015] S33. Find all unit numbers in At whose unit inspection time is not greater than tc z , place them in the array idy. If the length of the array idy ≥ m - 1, enter S34; otherwise, directly enter S36;

[0016] S34. After sorting the array pf idy in descending order, place the unit numbers corresponding to its largest first m - 1 elements in the array zy. zy is the group member unit number, and pf idy is the unit failure probability corresponding to all unit numbers in the array idy;

[0017] S35. Update s = s + 1, place [z zy] in the s-th row of matrix zall, and calculate the overall equivalent detection time;

[0018] S36. Update i = i + 1, if i ≤ na If yes, proceed to S32; otherwise, proceed to S4.

[0019] Preferably, the formula for calculating the overall equivalent detection time is as follows:

[0020]

[0021] Among them, tc z For the inspection time of the group leader unit, Pf z denoted as the failure probability of the group leader unit, and pg is the sum of the failure probabilities of all units except the group leader unit in the feasible unit partitioning result.

[0022] Preferably, the method further includes:

[0023] S6. Calculate the average troubleshooting time Tc using general-purpose testing tools:

[0024]

[0025] Among them, zPf j Let tx be the sum of the failure probabilities of all units checked in the j-th inspection. j pf is the cumulative inspection time of the previous j tests. r Let r be the probability of unit r failing.

[0026] Preferably, the formula for calculating the cumulative inspection time of the first j tests is as follows:

[0027]

[0028] Among them, ztc r This represents the maximum inspection time across all units during the r-th inspection.

[0029] Preferably, the method further includes:

[0030] S6. Calculate the probability distribution of fault diagnosis using general-purpose detection tools:

[0031]

[0032] Among them, px j It represents the probability of finding the faulty component in the first j iterations, and n represents the number of electronic units that make up the electronic device.

[0033] To achieve the above objectives, in a second aspect, the present invention provides a unit partitioning optimization system using a general detection tool, comprising: a processor and a memory; the memory for storing computer execution instructions; and the processor for executing the computer execution instructions, such that the method described in the first aspect is executed.

[0034] In summary, the technical solutions conceived by this invention have the following beneficial effects compared with the prior art:

[0035] This invention proposes a unit partitioning optimization method and system using a general detection tool. It constructs a candidate partitioning set by traversing the group leader unit and selecting the best member units. The group leader unit is the unit with the longest unit inspection time. When selecting the best member units, m-1 units with the highest probability of failure and no more than the group leader unit's state inspection time are selected from all currently uninspected units. These m-1 units, together with the feasible group leader units, constitute a feasible unit partitioning for the j-th detection. The partitioning with the minimum overall equivalent detection time is selected from all candidate feasible unit partitioning methods as the current partitioning, ultimately achieving optimized partitioning of all units. This partitioning optimization method achieves the globally optimal neighborhood solution, significantly reducing the average inspection time. Attached Figure Description

[0036] Figure 1 The present invention provides a flowchart of a unit partitioning optimization method using a general detection tool.

[0037] Figure 2 This is a schematic diagram illustrating the simulation verification results of the fault diagnosis time probability distribution provided 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] This invention provides a method for optimizing the division of electronic equipment into units using a general testing tool. This method can link the workload of troubleshooting, the time for troubleshooting, and the capacity, effectively reducing the average troubleshooting time and improving the timeliness of maintenance work.

[0040] Figure 1 This invention provides a flowchart of a cell partitioning optimization method using a general-purpose detection tool. (See flowchart for example.) Figure 1 As shown, the method includes:

[0041] Step S1. Obtain the capacity m of the general testing tool, the number of electronic units constituting the electronic device, the time for status testing of each electronic unit, and the probability of each electronic unit failing within the task time.

[0042] In engineering, commonly used electronic components are typically exponentially lifetime components, such as printed circuit board components, electronic parts, resistors, capacitors, and integrated circuits. Exponential cells are those whose lifetime follows an exponential distribution Exp(u), where the parameter u represents the mean lifetime. The density function of the exponential distribution is...

[0043] This invention stipulates that: (1) A certain piece of equipment is composed of multiple electronic units. For ease of description, the lifespan of each unit is described by time. (2) At any given time, at most one unit will fail. When a unit fails, it will affect the normal operation of the equipment, and the equipment will exhibit certain fault phenomena. At this time, repair work needs to be carried out. (3) When looking for the cause of the fault, the order of status checks on these units is independent and unrelated. That is, there is no situation where there is a specific requirement for the order of checks, such as "unit A must be checked first, and then unit B must be checked". (4) The lifespan distribution of each unit, the time consumed for (normal or not) status checks on each unit, and the time of the upcoming task are known. (5) After the detection tool starts checking multiple units simultaneously each time, the inspection results of each unit will only be given after all these units have been checked.

[0044] The relevant variables in this invention are defined as follows: the capacity of the detection tool is denoted as m; the number of units is denoted as n; the lifetime of unit i follows an exponential distribution Exp(u i The time consumed in checking the state of unit i is denoted as tc. i The task time is denoted as Tw. All these variables are known quantities.

[0045] This invention considers all units to be inspected in a single inspection using a testing tool as a group of units, where the unit that takes the longest to inspect is called the group leader unit, and the other units are called group member units.

[0046] The failure probability Pf of each unit is calculated by iterating through the units.

[0047] 1) Let i = 1;

[0048] 2) Integral calculation of Pf i :

[0049]

[0050] When k = i When k≠i

[0051] 3) Let i = i + 1. If i ≤ n, then execute 2); otherwise, execute S2.

[0052] Step S2. Round up the ratio of the number of electronic units to the capacity to get the maximum number of detections nm. Initialize the detection order j=1 and initialize the set of undetected units to all electronic units that make up the electronic device.

[0053] Let the maximum number of detections nm = n / m, and round the result up to the nearest integer. Place the cell numbers 1 to n into set A, and denote the number of cell numbers in set A as na.

[0054] Step S3. Find all feasible group leader units from the current unchecked unit set. The group leader unit is the unit with the longest unit inspection time. For each feasible group leader unit, select m-1 units with the highest failure probability and no more than the group leader unit status inspection time. Together with the feasible group leader units, they form a feasible unit partition for the j-th detection. All feasible unit partitions together form the feasible unit partition candidate set for the j-th detection.

[0055] Preferably, step S3 includes:

[0056] S31. Initialize the number of groups s = 0, i = 1, and the number of elements in the current undetected unit set na = n - (j - 1) * m, where n is the number of electronic units that constitute the electronic device;

[0057] S32. Initialize the group leader unit number z = A i The remaining elements in the undetected unit set A, excluding z, are candidate member unit numbers denoted as At;

[0058] S33. Find all cells in At and check if the time is no greater than tc. z The unit number is placed in the array idy. If the length of the array idy is greater than or equal to m-1, then proceed to S34; otherwise, proceed directly to S36.

[0059] S34. For array pf idy After sorting in descending order, the unit numbers corresponding to the largest m-1 elements are placed in the array zy, where zy is the unit number of the group member, and pf idy The probability of cell failure corresponding to all cells numbered in the array idy;

[0060] S35. Update s = s + 1, place [z zy] in the s-th row of matrix zall, and calculate the overall equivalent detection time;

[0061] S36. Update i = i + 1, if i ≤ na If yes, proceed to S32; otherwise, proceed to S4.

[0062] Preferably, the formula for calculating the overall equivalent detection time is as follows:

[0063]

[0064] Among them, tc z For the inspection time of the group leader unit, Pf z denoted as the failure probability of the group leader unit, and pg is the sum of the failure probabilities of all units except the group leader unit in the feasible unit partitioning result.

[0065] Step S4. Select the partitioning method with the minimum overall equivalent detection time from the feasible unit partitioning candidate set of the j-th detection as the unit partitioning result of the j-th detection, and update the undetected unit set to the difference between the current undetected unit set and the unit partitioning result of the j-th detection.

[0066] Step S5. Update j = j + 1. If j < nm, proceed to S3. Otherwise, use the current undetected unit set as the unit partitioning result of the last detection.

[0067] Preferably, the method further includes: S6. Calculating the average troubleshooting time Tc detected using a general-purpose testing tool:

[0068]

[0069] Among them, zPf j Let tx be the sum of the failure probabilities of all units checked in the j-th inspection. j pf is the cumulative inspection time of the previous j tests. r Let r be the probability of unit r failing.

[0070] Preferably, the formula for calculating the cumulative inspection time of the first j tests is as follows:

[0071]

[0072] Among them, ztc r This represents the maximum inspection time across all units during the r-th inspection.

[0073] Preferably, the method further includes: S6. Calculating the fault diagnosis probability distribution detected using a general detection tool:

[0074]

[0075] Among them, px j It represents the probability of finding the faulty component in the first j iterations, and n represents the number of electronic units that make up the electronic device.

[0076] The time consumption result of evaluating the unit division result Mdy.

[0077] 1) Initialize the detection sequence number j = 1, and the average fault troubleshooting time Tc = 0.

[0078] 2) Place the non-zero element of the j-th row in matrix Mdy into array zj, where zj stores the unit number of the j-th detection.

[0079] 3) Find the maximum check time for the corresponding cell from all cells in zj and save it to ztc. j Accumulate the failure probabilities of all units in zj and save them to zPf. j .

[0080] 4) Update Where pxj is in time tx j The probability of finding the faulty component inside.

[0081] 5) Update j = j + 1. If j ≤ nm, then execute 2). Otherwise, terminate the calculation and output the unit partitioning result Mdy, average fault troubleshooting time Tc, fault troubleshooting time probability distribution tx and px for each detection when using the general detection tool.

[0082] Example

[0083] A component consists of 20 electronic units, with a task duration of 100 hours and a testing tool capacity of 5. Relevant information is shown in Table 1. Using the method described above, calculate the average fault detection time under these conditions.

[0084] Table 1

[0085] Unit Number Distribution parameter u Status check time / min 1 865 7 2 1170 55 3 3705 48 4 670 8 5 2350 12 6 2940 25 7 2480 14 8 3935 28 9 1400 35 10 2325 44 11 1700 52 12 3990 22 13 2240 59 14 2330 47 15 2490 35 16 1105 43 17 3165 44 18 4025 45 19 810 61 20 1855 54

[0086] 1) Calculate the probability Pf of each unit failing, and the results are shown in Table 2.

[0087] Table 2

[0088]

[0089]

[0090] 2) Let the maximum number of detections nm = 4, place the unit numbers 1 to 20 in set A, the number of unit numbers in set A is na = 20, and the detection sequence number j = 1.

[0091] 3) Optimize the calculation of the feasible cell partition zall and its equivalent detection time tpf for the j-th detection. When j=1, the cell partition zall and its equivalent detection time tpf are shown in Table 3.

[0092] Table 3

[0093]

[0094] 4) Based on tpf, select the optimal cell from zall and partition it, and save the result to Mdy.

[0095] The minimum value tpm = 100.12 in the array tpf is denoted as im = 8. The element in the im-th row of the matrix zall is selected as the unit index zj = [12, 4, 1, 5, 7] of the first detection and stored in the first row of the matrix Mdy.

[0096] 5) Update j = j + 1, delete zj from A, na = na - m, i = 1. If j < nm, execute 3). Otherwise, directly place the elements in A into the nm-th row of matrix Mdy.

[0097] Execute steps 3) and 4) multiple times to optimize the division of the detection at j=2, 3, and 4 in sequence. The Mdy calculation results are shown in Table 4.

[0098] Table 4

[0099]

[0100] 6) Evaluate the time consumption of the unit division result Mdy, Tc = 87.2, and the probability distribution result px of the fault troubleshooting completion time tx is shown in Table 5.

[0101] Table 5

[0102]

[0103] 7) Terminate the calculation and output the unit partitioning result Mdy, the average fault troubleshooting time Tc, and the probability distribution px of the fault troubleshooting time tx. The optimized fault troubleshooting scheme is as follows: The unit numbers for the first inspection are 12, 4, 1, 5, and 7, which takes 22 minutes and has a probability of finding the faulty component of 0.32. If the unit numbers for the first inspection are normal, the second, third, and fourth inspections are performed sequentially. The unit numbers for the second inspection are 19, 16, 2, 9, and 11; the unit numbers for the third inspection are 14, 10, 15, 6, and 17; and the unit numbers for the fourth inspection are 3, 8, 13, 18, and 20. The probabilities of finding the faulty component within 83, 130, and 189 minutes are 0.69, 0.85, and 1.00, respectively. The average fault troubleshooting time is 87.2 minutes.

[0104] A simulation model can be established to verify the correctness of the above method. The simulation model is briefly described as follows:

[0105] (1) Generate n random numbers simT i , 1≤i≤n, simT i It follows the lifetime distribution law of unit i.

[0106] (2) In all simT iFind the smallest number in the range, and denote the corresponding index as g, i.e.: simT g ≤simT i , 1≤i≤n.

[0107] (3) If simT g If <Tw is true, then this simulation is valid. According to the detection order, use the detection tool to detect up to m units each time until unit g is detected. The detection time for finding the cause of the fault can be obtained.

[0108] After numerous simulations, the average troubleshooting time and its probability distribution can be statistically obtained.

[0109] Figure 1 The figure shows the probability distribution results of the fault diagnosis time for the above examples obtained by simulation and the method of this invention, respectively. As can be seen from the figure, the results of the two methods are in excellent agreement. For the above examples, 10,000 fault diagnosis schemes were randomly generated. The simulation results for the minimum, maximum, and average fault diagnosis times were 101.2, 168.8, and 137.8 minutes, respectively. Compared with the result of 87.2 minutes obtained by the method of this invention, the optimization effect of the method of this invention is significant.

[0110] Table 6 shows the simulation results of the average fault diagnosis time for the above examples and the evaluation results of the method of the present invention when the detection tool capacity is 1 to 10. As can be seen from Table 6, the two are highly consistent.

[0111] Table 6

[0112]

[0113] Using the method of this invention, by traversing the average detection time of each detection tool capacity within a certain range, results similar to Table 6 can be obtained, which can be used to assist in selecting the detection tool capacity. For example, as can be seen from Table 6, when the capacity exceeds 4, the reduction in average troubleshooting time is very limited. At this time, from the perspective of cost-effectiveness, the capacity should not exceed 4. If the average troubleshooting time is required to be no more than 100 minutes, then the capacity should be at least 4.

[0114] Those skilled in the art will readily understand that the above description is merely 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 scope of protection of the present invention.

Claims

1. A method for optimizing cell partitioning using a general-purpose detection tool, characterized in that, The method includes: S1. Obtain the capacity m of the general detection tool, the number of electronic units constituting the electronic device, the time for detecting the status of each electronic unit, and the probability of each electronic unit failing during the task time. S2. Round up the ratio of the number of electronic units to the capacity as the maximum number of detections nm, initialize the detection order j = 1, and initialize the set of undetected units as all the electronic units constituting the electronic device. S3. Find all feasible group leader units from the current set of undetected units, where the group leader unit is the unit with the longest unit inspection time; for each feasible group leader unit, select no more than m - 1 units with the highest failure probability and the longest status inspection time of the group leader unit, and jointly form a feasible unit partition for the j-th detection with the feasible group leader unit. All feasible unit partitions jointly form the candidate set of feasible unit partitions for the j-th detection. S4. Select the partition method with the minimum overall equivalent detection time from the candidate set of feasible unit partitions for the j-th detection as the unit partition result for the j-th detection, and update the set of undetected units as the difference set between the current set of undetected units and the unit partition result of the j-th detection. S5. Update j = j + 1. If j < nm, enter S3; otherwise, use the current set of undetected units as the unit partition result for the last detection. S6. Calculate the average troubleshooting time using general-purpose testing tools. : in, Let be the sum of the failure probabilities of all units during the j-th inspection. This represents the cumulative inspection time of the previous j tests. Let r be the probability of unit r failing, and n be the number of electronic units that make up the electronic device. Among them, step S3 includes: S31. Initialize the number of groups , The number of elements in the current undetected unit set, na = n - (j - 1), is given by equation na. m; S32. Initialize the group leader unit number The remaining elements in the undetected unit set A, excluding z, are candidate member unit numbers denoted as At; S33. Find all cells in At and check if the time is no greater than [time not specified]. The unit number is placed in an array. In the array, if If the length is greater than or equal to m-1, proceed to S34; otherwise, proceed directly to S36. S34. For arrays After sorting in descending order, place the cell numbers corresponding to the largest m-1 elements into an array. In, among them, Assign unit numbers to group members. Number all cells in the array The probability of unit failure corresponding to the unit in the diagram; S35 Update ,Bundle Placed in the first position of matrix zall Okay, and calculate the overall equivalent detection time; S36. Update ,like If yes, proceed to S32; otherwise, proceed to S4. The overall equivalent detection time The calculation formula is as follows: in, The inspection time for the group leader unit. The failure probability of the group leader unit. This is the sum of the failure probabilities of all units except the group leader unit in the feasible unit partitioning result.

2. The method as described in claim 1, characterized in that, Cumulative examination time of the previous j tests The calculation formula is as follows: in, This represents the maximum inspection time across all units during the r-th inspection.

3. The method as described in claim 1, characterized in that, The method further includes: calculating the fault troubleshooting probability distribution detected by the general detection tool: in, It represents the probability of finding the faulty component in the first j iterations.

4. A cell partitioning optimization system using a general-purpose detection tool, characterized in that, including: a processor and a memory; The memory is used to store computer execution instructions; the processor is used to execute the computer execution instructions so that the method according to any one of claims 1 to 3 is executed.

Images