WO2022271171A1

WO2022271171A1 - System and method to determine financial incentive based on maintenance activity

Info

Publication number: WO2022271171A1
Application number: PCT/US2021/038745
Authority: WO
Inventors: Yusuke Jin
Original assignee: Hitachi, Ltd.
Priority date: 2021-06-23
Filing date: 2021-06-23
Publication date: 2022-12-29

Abstract

Example implementations are directed to determining incentives based on maintenance execution conditions corresponding to the recommended maintenance schedule. The implementations involve a function recommending a maintenance schedule for the asset by predicting remaining useful lifetime and/or detecting failure of insured assets, a function for evaluating actual maintenance timing by comparing with the recommended maintenance schedule, a function scoring trust in view of ability to flexibly respond to defects and/or in view of ability to accurately determine the need for maintenance, and a function calculating insurance premium discount based on the trust score. The trust score calculation is not only based on the analysis result of the underlying machinery, but also maintenance activity. Such evaluation is based on flexibility of urgent maintenance recommendation and judgement ability to accepting recommended maintenance. The trust score is used to calculate insurance premium discount or other financial incentives.

Description

SYSTEM AND METHOD TO DETERMINE FINANCIAL INCENTIVE BASED ON

MAINTENANCE ACTIVITY

BACKGROUND

Field

[0001] The present disclosure is generally directed to maintenance systems, and more specifically, to systems and methods to determine financial incentives from maintenance activity.

Related art

[0002] In the related art implementations, there are Internet of Things (IoT) apparatus credibility calculation systems, devices, and methods, that acquire measurement data emitted over time from an IoT apparatus having a sensor, acquire non-measurement information and calculate the credibility of the measurement data and the non-measurement information.

SUMMARY

[0003] In such related art implementations, there are no implementations that calculate the trust score or credibility of the insured party based on maintenance related activity. There are also no implementations involving calculating the trust score based on whether the maintenance execution was prior to the recommended maintenance schedule.

[0004] Example implementations can involve systems and methods to determine a financial incentive, such as insurance premium discount, based on a maintenance execution condition corresponding to recommended maintenance schedule. The method includes a function recommending maintenance schedule for the insured asset by predicting remaining useful lifetime and/or detecting failure of insured assets. The method also includes a function evaluating actual maintenance timing by comparing with the recommended maintenance schedule. The method also includes a function scoring trust of the insured in view of ability to flexibly respond to defects and/or in view of ability to accurately determine the need for maintenance. The method also includes a function calculating insurance premium discount based on the insured’s trust score. [0005] Example implementations involve a calculation method of the trust score of the insured party as based not only from an analysis result of the underlying machinery, but also the maintenance activity undertaken by the insured party. Such evaluation is based on flexibility of urgent maintenance recommendation and judgement ability to accepting recommended maintenance. The trust score is used to calculate insurance premium discount or other financial incentives.

[0006] Aspects of the present disclosure can involve a computer program having instructions for tracking of maintenance for a machine in a machine operator environment, the instructions involving processing timing information of the machine operator environment from receipt of a maintenance scheduled time indicative of a scheduled timing of the maintenance for the machine and a maintenance execution time indicative of an actual timing of execution of the maintenance for the machine; determining acceptance of a maintenance recommendation based on comparisons between each of the maintenance scheduled time, the maintenance execution time, and a maintenance recommended time indicative of a recommended maintenance timing of the maintenance for the machine; and calculating a trust score for the machine operator environment based on the acceptance of the maintenance recommendation. The computer program and instructions may be stored in a non-transitory computer readable medium and executed by one or more processors.

[0007] Aspects of the present disclosure can involve a method for tracking of maintenance for a machine in a machine operator environment, the method involving processing timing information of the machine operator environment from receipt of a maintenance scheduled time indicative of a scheduled timing of the maintenance for the machine and a maintenance execution time indicative of an actual timing of execution of the maintenance for the machine; determining acceptance of a maintenance recommendation based on comparisons between each of the maintenance scheduled time, the maintenance execution time, and a maintenance recommended time indicative of a recommended maintenance timing of the maintenance for the machine; and calculating a trust score for the machine operator environment based on the acceptance of the maintenance recommendation.

[0008] Aspects of the present disclosure can involve a system for tracking of maintenance for a machine in a machine operator environment, the system involving means for processing timing information of the machine operator environment from receipt of a maintenance scheduled time indicative of a scheduled timing of the maintenance for the machine and a maintenance execution time indicative of an actual timing of execution of the maintenance for the machine; means for determining acceptance of a maintenance recommendation based on comparisons between each of the maintenance scheduled time, the maintenance execution time, and a maintenance recommended time indicative of a recommended maintenance timing of the maintenance for the machine; and means for calculating a trust score for the machine operator environment based on the acceptance of the maintenance recommendation.

BRIEF DESCRIPTION OF DRAWINGS

[0009] FIG. 1 illustrates an example system architecture of an incentive calculation system, in accordance with an example implementation.

[0010] FIG. 2 illustrates an example physical configuration of an incentive calculation node, in accordance with an example implementation.

[0011] FIG. 3 illustrates an example conceptual diagram of an incentive calculation system effect in incentive calculator, in accordance with an example implementation.

[0012] FIG. 4 is an example flow diagram illustrating an example process of an incentive calculator, in accordance with an example implementation.

[0013] FIG. 5 is a flow diagram illustrating an example process of maintenance timing evaluator, in accordance with an example implementation.

[0014] FIG. 6 is a flow diagram illustrating an example process of trust score updater, in accordance with an example implementation.

[0015] FIG. 7 is a data structure illustrating an example information of trust score updating table, in accordance with an example implementation.

[0016] FIG. 8 illustrates an example conceptual diagram of the judgement ability evaluation of the trust score updater, in accordance with an example implementation.

[0017] FIG. 9 is a data structure illustrating example information for training data, in accordance with an example implementation.

PET ATT /ED DESCRIPTION [0018] The following detailed description provides details of the figures and embodiments of the present application. Reference numerals and descriptions of redundant elements between figures are omitted for clarity. Terms used throughout the description are provided as examples and are not intended to be limiting. For example, the use of the term “automatic” may involve fully automatic or semi-automatic implementations involving user or administrator control over certain aspects of the implementation, depending on the desired implementation of one of ordinary skill in the art practicing implementations of the present application. Selection can be conducted by a user through a user interface or other input means, or can be implemented through a desired algorithm. Embodiments as described herein can be utilized either singularly or in combination and the functionality of the embodiments can be implemented through any means according to the desired implementations.

[0019] The example implementations described herein are provided with the example of a factory environment. However, the example implementations described herein may also be extended to other machine operator environments besides the factory environment, and the present disclosure is not limited thereto. For example, the machine operator environments can include fleet management (e.g., truck) or field machinery operation (e.g., construction, mining) environments.

[0020] Financial resilience can be an important factor for manufacturers. Predictive maintenance technology is introduced to increase maintenance efficiency and accuracy through gathering machine operation or factory environmental data. If the predictive maintenance works well, then the chances of factory operation stoppage is reduced, thereby saving on insurance payments. On the other hand, insurance companies are paying attention to utilizing these industrial field data and develop IoT insurance to realize a usage-based insurance product.

[0021] It is important for the manufacturer to follow the maintenance recommendation based on the IoT data analysis in order to benefit from the predictive maintenance service. Thus, reliability of the adoption of maintenance recommendations by the insured party is an issue for determining the appropriate insurance premium. However, determining the reliability correctly from the actions of the insured party is an issue that is not addressed by related art systems. Further, the insured party should not necessarily be penalized from failing to adopt unnecessary maintenance recommendations, as the proposed maintenance recommendations by the insurer may not be applicable to the insured party or may not be correct. [0022] To address the above issues in the related art, the example implementations involve a solution having systems and methods for determining financial incentive such as insurance premium discount based on maintenance execution condition corresponding to recommended maintenance schedule.

[0023] The method includes a function recommending a maintenance schedule for the insured asset by predicting remaining useful lifetime and/or detecting failure of insured assets. The example implementations also include a function evaluating actual maintenance timing by comparing with the recommended maintenance schedule. The example implementations also include a function scoring trust of the insured in view of ability to flexibly respond to defects and/or in view of ability to accurately determine the need for maintenance. The example implementations also include a function calculating insurance premium discount based on the trust score of the insured.

[0024] If the accuracy of maintenance recommendations is high, then the financial incentives such as the insurance premium discount can be decided based on the response to the maintenance recommendation. However, there is an asymmetric information problem in insurance. For example, the life insurance provider cannot determine the risk correctly if the insured party did not provide the appropriate health conditions to the life insurance provider. In the field of industrial insurance, if certain important field data cannot be provided to the insurer, then the prediction result will be inaccurate. Therefore, sometimes it may be reasonable for the insured party not to accept a maintenance recommendation from the insurer.

[0025] Therefore, in example implementations described herein, the trust score of the insured may also be based on actual loss tendency in conjunction with acceptance of maintenance recommendations so as to not penalize the insured party if the insured party utilizes its own judgment to not execute a maintenance recommendation and no actual loss occurs as a result.

[0026] In example implementations described herein, the trust score of the insured is calculated not only from the machinery analysis result, but also from the maintenance activity from the insured. Such an evaluation is based on the flexibility of an urgent maintenance recommendation and the judgement ability from accepting the recommended maintenance. The trust score is used to calculate an insurance premium discount or other financial incentives. [0027] Example implementations described herein involve systems and methods to determine financial incentive based on maintenance activity works.

[0028] FIG. 1 illustrates an example system architecture of an incentive calculation system, in accordance with an example implementation. System 101 has insurer node(s) 102, insured factory node(s) 103 and incentive calculation node(s) 104. All components are connected through a network 105. Insured factory node 103 receives trust score information 215 from incentive calculation node 104, which is used to calculate the insurance premium based on the trust score of the insured calculated in the incentive calculation node 104. Insured factory node 103 provides operational log information 209, environmental log information 210, failure information 211, maintenance execution information 212 and maintenance scheduling information 213 to incentive calculation node 104. Incentive calculation node 104 is used to recommend the maintenance schedule for the asset of the insured, evaluate the actual maintenance timing, calculate the trust score of the insured and provide the score to insured factory node 102 for the insurance premium discount calculation. All components are connected though network 105. Depending on the desired implementation, insurer node(s) 102 and incentive calculation node(s) 104 can be integrated as the same entity.

[0029] FIG. 2 illustrates an example physical configuration of incentive calculation node 104, in accordance with an example implementation. Incentive calculation node 104 can include memory 201, local storage 202, communication interface(s) 203, processor(s) 204 and I/O Devices(s) 205. Local storage 202 contains operating system 206, maintenance recommender 207, incentive calculator 208, maintenance timing evaluator 216, trust score updater 217, machine leaning module 219, inference module 220, operational log information 209, environmental log information 210, failure information 211, maintenance execution information 212, maintenance scheduling information 213, maintenance recommendation information 214, trust score information 215 trust score updating table 218, insurance data 221, training data 222 and model parameter 223.

[0030] Maintenance recommender 207 is software application providing function for recommending maintenance schedule for the insured asset by predicting remaining useful lifetime and/or detecting failure of insured assets. Maintenance recommender 207 reads input data stored in operational log information 209, environmental information 210, failure information 211 and maintenance execution information 212. Maintenance recommender 207 predicts when to execute next maintenance by analyzing remaining useful lifetime (RUL) and/or detecting anomaly of machinery, then stores the recommended maintenance schedule to maintenance recommendation information 214. The implementation of determining RUL, failure detection, and maintenance recommendation can be in accordance with any desired implementation known to one of ordinary skill in the art and is not particularly limited.

[0031] Incentive calculator 208 is a software application providing a function for evaluating actual maintenance timing based on a comparison with the recommended maintenance schedule, and for providing a trust score of the insured based on the ability to flexibly respond to defects and/or the ability to accurately determine the need for maintenance, as well as for calculating the insurance premium discount rate based on the trust score of the insured. Incentive calculator 208 reads input data stored in maintenance execution information 212, maintenance scheduling information 213 and maintenance recommendation information 214, and stores the calculation result to trust score information 215. The detailed flow diagram of incentive calculator 208 is described in FIG. 4.

[0032] Maintenance timing evaluator 216 and trust score updater 217 are software applications called from incentive calculator 208 and described in FIG. 5 and FIG. 6.

[0033] Machine leaning module 219 is a software application learning from training data 222 as described herein and outputting model parameter 223 based on defined hyper parameter which is tuned with a searching method such as Grid Search, Random Search, and/or Bayes Search.

[0034] Inference module 220 is a software application that infers a certain result from input data using model parameter 223.

[0035] Operational log information 209 is a data store which may include a customer identifier (ID) for identifying the insured customer, asset ID for identifying the insured asset, asset name, asset type, and manufacturing operation information (e.g., product name, product type, process type, product size or attached position, and so on).

[0036] Environmental information 210 is a data store which may include a customer ID for identifying the insured customer, asset ID for identifying the insured asset, and environmental information around the operation asset such as temperature, moisture shock, and so on. [0037] Failure information 211 is a data store which may include a customer ID for identifying the insured customer, asset ID for identifying the insured asset, failure ID for identifying the occurred failure, and start and end time of an operation outage.

[0038] Maintenance execution information 212 is a data store which may include a customer ID for identifying the insured customer, asset ID for identifying the insured asset, maintenance ID for identifying an executed maintenance operation, type and description of a maintenance operation, maintenance operator, and maintenance execution time (Te) 305 as will be described herein.

[0039] Maintenance scheduling information 213 is a data store which may include a customer ID for identifying the insured customer, asset ID for identifying the insured asset, scheduling ID for identifying scheduled maintenance, scheduled maintenance operation, and maintenance scheduled time (Ts) 303 as will be described herein.

[0040] Maintenance recommendation information 214 is a data store which may include a customer ID for identifying the insured customer, asset ID for identifying the insured asset, recommendation ID for identifying maintenance recommendation, recommended maintenance operation and maintenance recommendation time (Tr) 304 as will be described herein.

[0041] Maintenance recommendation information 214 also includes a recommendation acceptance flag, which indicates whether the recommendation has been accepted or not. The flag may be one of ACCEPTED, PARTIALLY ACCEPTED and NOT ACCEPTED. This flag will be set in Step 506, 507 and 508 of FIG. 5.

[0042] Trust score information 215 is a data store which may include a customer ID for identifying the insured customer and the trust score of the customer. The trust score is a score associated with the insured that is indicative of the maintenance activity of the insured party. Such a trust score can be utilized or normalized with desired metrics to determine aspects such as the insurance premium rate, loan percentage rate, and so on. For example, trust score can be defined as a value between 100 and 900 with a range of over 750 being categorized as VERY GOOD, which can then be normalized to corresponding insurance premium rates, loan percentage rates, rebates, and so on, but is not limited thereto.

[0043] Trust score updating table 218 is data store for determine increase and decrease of trust score. The detail of this data store is shown in FIG. 8. [0044] Insurance data 221 is a data store which may include insurance premium, insured amount and other coverage information related to the policy of the insured received from insurer node. It may be used as part of training data 222.

[0045] Training data 222 is data store is input data for machine learning module 219. The detail of this data store is shown in FIG. 9.

[0046] Model parameter 223 is parameter data for inference module to infer the result from input data and is trained by machine learning module 219.

[0047] FIG. 3 illustrates an example conceptual diagram of incentive calculation system effect in incentive calculator 208, in accordance with an example implementation. Maintenance scheduled time (Ts) 303 is the scheduled timing of a machine maintenance. Maintenance recommendation time (Tr) 304 is the recommended maintenance timing of the relevant machine by maintenance recommender 207. Maintenance execution time (Te) 305 is actual timing of maintenance execution of the relevant machine after the recommendation. Recommended advance time 306 is the difference between Tr 304 and Ts 303. Actual advanced time 307 is the difference between Te 305 and Ts 303.

[0048] Case-1 301 shows the case that the insured follows the recommendation. In this case, actual advanced time 307 is longer than recommended advance time 306, in other words, Te < Tr. Case-2 302 shows the case that the insured partially follows the recommendation and it cannot advance the maintenance execution. In this case, actual advanced time 307 is shorter than recommended advance time 306, in other words, Tr > Te. Furthermore, there could be also Case-3 in which maintenance will be executed as scheduled or later, in other words, Ts <Te.

[0049] Depending on the desired implementation, the trust score can be increased or decreased based on the case of the executed maintenance.

[0050] There is a multiple adjustment pattern in response to the strategy of the insurer. For example, the insurer can select maintenance flexibility as an evaluation criteria. In this pattern, the flexible decision making to accept a recommendation requiring an earlier maintenance operation than expected is factored into the trust factor. For example, such a decision can cause an adjustment such that the score increases in Case-1, increases based on the advance period in Case-2 and decreases in Case-3. Further details are provided herein. [0051] FIG. 4 is an example flow diagram illustrating an example process of incentive calculator 208, in accordance with an example implementation.

[0052] The flow begins at 401, which can be started voluntarily or regularly, depending on the desired implementation. At 402, the incentive calculator 208 receives up-to-date information including maintenance scheduled time (Ts) 303 and maintenance execution time (Te) 305 of a machine from incentive calculation node 104 through network 105. Then incentive calculator 208 stores maintenance scheduled time (Ts) 303 into maintenance scheduling information 213 and maintenance execution time (Te) 305 of the relevant machine into maintenance execution information 212.

[0053] At 403, the incentive calculator 208 compares maintenance scheduled time (Ts) 303, maintenance recommendation time (Tr) 304 and maintenance execution time (Te) 305 and their time order. The detailed procedure of this step is described with respect to FIG. 5.

[0054] At 404, the incentive calculator 208 calculates the trust score of the insured based on the order clarified in Step 403. The detailed procedure of this step is described with respect to FIG. 6.

[0055] At 405, the incentive calculator 208 calculates insurance premium discount rate based on the trust score of the insured. Instead of the discount rate of insurance, it can apply to calculation of preferential interest rate of loan or other financial incentives.

[0056] At 406, the incentive calculator 208 quits the process.

[0057] FIG. 5 is a flow diagram illustrating an example process of maintenance timing evaluator 216, in accordance with an example implementation. The flow begins at 501.

[0058] At 502, the maintenance timing evaluator 216 get maintenance scheduled time (Ts)

303 from maintenance scheduling information 213, maintenance recommendation time (Tr)

304 from maintenance recommendation information 214 and maintenance execution time (Te)

305 from maintenance execution information 212.

[0059] At 503, the maintenance timing evaluator 216 compares Te 305 with Ts 303. If Te is less than or equal to Ts (503: Yes), then the maintenance timing evaluator 216 proceeds to Step 504. If not (No), proceeds to Step 509. [0060] At 504, the maintenance timing evaluator 216 compares Te 305 with Tr 304. If Te is less than Tr (504: Yes), maintenance timing evaluator 216 proceeds to Step 506. If not (No), then the process proceeds to Step 505.

[0061] At 505, the maintenance timing evaluator 216 compares Te 305 with Ts 303. If Te is equal to Ts (505: Yes), maintenance timing evaluator 216 proceeds to Step 508. If not (No), proceeds to Step 507.

[0062] At 506, the maintenance timing evaluator 216 sets the recommendation acceptance flag to ACCEPTED in maintenance recommendation information 214.

[0063] At 507, the maintenance timing evaluator 216 sets the recommendation acceptance flag to PARTIALLY ACCEPTED in maintenance recommendation information 214.

[0064] At 508, the maintenance timing evaluator 216 sets the recommendation acceptance flag to NOT ACCEPTED in maintenance recommendation information 214.

[0065] At 509, the maintenance timing evaluator 216 quits the process.

[0066] FIG. 6 is a flow diagram illustrating an example process of trust score updater 217, in accordance with an example implementation. The flow begins at 601.

[0067] At 602, the trust score updater 217 calculates the trust score of the insured based on the flexibility of the maintenance execution. In this step, the trust score updater 217 evaluates an ability for flexible execution of a recommended maintenance. Companies which can cope with an urgent maintenance requirement are deemed to be trustworthy in maintenance operations and therefore have a lower risk of failure and production outage. Then, insurance companies can provide a premium discount or preferential interest rate loan to such companies.

[0068] Trust score updater 217 can evaluate each maintenance operation and update the trust score or evaluate maintenance operations for a certain period. In the latter case, trust score updater 217 aggregates these maintenance operations, and if the number of times of recommendation acceptance exceeds a defined threshold, then the trust score is increased. Further, the trust score updater 217 can suspend to decrease trust score when the number of times of recommendation acceptance is less than a threshold. Example of the increase and decrease patterns above are defined in trust score updating table, described in FIG. 7. [0069] Additionally, depending on the desired implementation, trust score updater can also be configured to not increase the trust score (e.g., maintain the same score or decrease the score based on the desired implementation), for the insured party if the maintenance recommendation was only partially accepted so as to encourage the insured party to fully accept the maintenance recommendation. Such a situation can occur, for example, when the maintenance is executed later than the maintenance recommendation time, but before the maintenance scheduled time.

[0070] At 603, the trust score updater 217 calculates the trust score of the insured based on the judgement of the maintenance execution. In this step, trust score updater 217 evaluates the ability to judge the need of executing a recommended maintenance. For example, suppose the insured schedules a maintenance operation based on their own operational data, which is not provided to insurers. As a result, the decision by the insured party not to accept the recommendation is sometimes correct or reasonable. So, trust score updater 217 calculates the trust score as follows. If the insured did not execute the recommended maintenance and a failure subsequently occurs, then the trust score updater 217 decreases the trust score. If the insured did not execute the recommended maintenance and failure has not subsequently occurred yet, then the trust score updater 217 increases the trust score because the decision of the insured party turned out to be correct, and it may have ability to make the right decision about maintenance.

[0071] This step can be processed using artificial intelligence. A detailed example using machine learning is described as follows.

[0072] Prior to this step, machine learning module 219 learns from training data 222 and has created model parameter of inference module 220, which infers the expected value of the loss amount caused by the decision for acceptance of recommended maintenance. The training data 222 includes recommendation information such as recommendation timing, insured asset type, product type of the insured asset and acceptance condition calculated in step 506, 507 and 508. The training data 222 is described below and shown in FIG. 9.

[0073] In this step, inference module 220 inputs certain recommendation information and its acceptance condition, and infers an expected value of loss amount which is caused by the acceptance decision. If there are multiple recommendations to the insured company, the expected value is determined and the expected profit of insurance company is calculated by comparing the total expected loss amount with the insurance premium. In case that the expected profit increases, trust score updater 217 increases the trust score of the insured. (For example, claimed maintenance cost is reduced by not accepting recommended maintenance). In case that the expected profit decreases, trust score updater 217 decreases trust score of the insured. The increase/decrease range of the trust score could be defined as a table, or otherwise in accordance with the desired implementation.

[0074] An example conceptual diagram is shown in FIG. 8 and an example of training data is shown in FIG. 9.

[0075] At 604, the trust score updater 217 calculates the trust score of the insured by combining the result of Step 602 and 603. Trust score updater 217 can adjust the weighing of the result. For example, when the insurer has gathered enough data for a precise recommendation, the result of the flexibility evaluation should be a higher weight. On the other hand, if the data gathered is insufficient, then the result of the judgement ability evaluation of the insured should be high weight. Some weighted table for combining such results can be used in this step, or otherwise in accordance with the desired implementation.

[0076] In this example implementation, two evaluation indicators as shown in Step 602 and 603 is used. However, more than three evaluation methods can be used and can be combined to calculated trust score, and the present disclosure is not limited thereto.

[0077] At 605, the trust score updater 217 quits the process.

[0078] FIG. 7 is a data structure illustrating an example information of trust score updating table 218, in accordance with an example implementation. The information can include scoring pattern ID 701, aggregation unit 702, scoring condition 703 and score increase / decrease 704. Scoring pattern ID 701 is an identifier for each trust score updating pattern. Aggregation unit 702 is a unit to aggregate maintenance operation before the score calculation. For example, if the data of aggregation unit 702 column is “Each”, then the trust score is calculated for each maintenance operation.

[0079] Scoring condition 703 is a calculation condition of each score updating pattern and score increase / decrease 704 is how the score increases or decreases when the condition is met. For example, row R003 indicates that data of aggregation unit 702 column shows two weeks, data of scoring condition 703 column shows ACCEPTED > 50% and data of score increase / decrease 704 column shows +15. Row R003 indicates that if 50% of the recommended maintenance operation is executed in the weeks, then the trust score increases by 15 points.

[0080] FIG. 8 illustrates an example conceptual diagram of the judgement ability evaluation in Step 603 of trust score updater 217, in accordance with an example implementation.

[0081] Average advanced time 802 (horizontal axis) shows an average difference of time between maintenance execution time (Te) 305 and maintenance recommendation time (Tr) 304. Additional loss 803 (vertical axis) shows the total loss for a certain period, which includes additional maintenance fees due to accepting the recommended maintenance and failure loss incurred from not accepting the recommended maintenance.

[0082] Each plot stands for an insured company. Group 1 (plots around 804) had a very large loss, because these companies tend not to accept recommended maintenance and failures often occur as a result. Group 2 (plots around 805) had a large loss, because these companies tend to easily accept recommendations which may contain unnecessary maintenance. On the other hand, Group 3 (plots around 806) indicates low additional loss and tends to accept recommendation appropriately. Based on the example results, Group 3 (806) seems to have the judgement ability of accepting recommended maintenance appropriately. Thus, trust score updater 217 can give such companies a higher trust score.

[0083] Furthermore, such an evaluation axis can be replaced with another indicator, such as environmental impact, or otherwise in accordance with the desired implementation. Then, the trust score can be used for Environmental, Social, and Corporate Governance (ESG) investment or other sustainable finance scheme.

[0084] To determine the groups and the boundaries of the groups for classifying factory environments based on their maintenance timing information, various machine learning algorithms that utilize clustering based on the historical maintenance timing information obtained from other factory environments as illustrated in FIG. 9 can be used to formulate the groups illustrated in FIG. 8 and train the machine learning algorithm to classify other factory environments accordingly. For example, such machine learning algorithms can involve K- Means Clustering, Mean-Shift Clustering, Density-Based Spatial Clustering of Applications with Noise, Expectation-Maximization Clustering using Gaussian Mixture Models, Hierarchical Clustering, and so on, but is not limited thereto. [0085] FIG. 9 is a data structure illustrating example information for training data 219, in accordance with an example implementation. Training data 219 can include recommendation ID 901, recommended maintenance timing 902, asset type 903, product type 904, process type 905, scoring condition 906 and loss amount 907, but is not limited thereto and can omit or include information in accordance with the desired implementation. For example, other information for training data 219 can include a Remaining Useful Life (RUL) analytics algorithm type which was used to recommendation, production line layout information, and catastrophe risk information.

[0086] Recommendation ID 901 indicates an identifier for each recommended maintenance. Maintenance recommendation time 902 indicates the recommended maintenance timing of the relevant machine. Asset type 903 indicates the type of insured asset corresponding to the recommendation in recommendation ID 901. Product type 904 indicates the type of product which is produced with the corresponding asset in asset type 903. Process type 905 indicates the type of production process which is used to produce the product corresponding product in product type 904. Scoring condition 906 is result of maintenance timing evaluator 216 which is shown in FIG. 5.

[0087] Loss amount 907 stands for actual loss which has occurred and is due to the insurance company. It could be the claim amount from the insured corresponding to the failed machinery. If policies of the insured include coverage for the maintenance cost corresponding to the recommended maintenance, the loss amount 907 could include maintenance cost. If policies of the insured include coverage for loss of profit, loss amount 907 could include profit loss from production line halting caused by executing the recommended maintenance.

[0088] Recommendation ID 901, maintenance recommendation time 902 and scoring condition 906 could be copied from maintenance recommendation information 214. Asset type 903, product type 904 and process type 905 could be copied from operational log information 209. Loss amount 907 could be copied or calculated from insurance data 221.

[0089] In a first aspect, there is a method for tracking of maintenance for a machine in a machine operator environment, the method comprising processing timing information of the machine operator environment from receipt of a maintenance scheduled time indicative of a scheduled timing of the maintenance for the machine and a maintenance execution time indicative of an actual timing of execution of the maintenance for the machine; determining acceptance of a maintenance recommendation based on comparisons between each of the maintenance scheduled time, the maintenance execution time, and a maintenance recommended time indicative of a recommended maintenance timing of the maintenance for the machine; and calculating a trust score for the machine operator environment based on the acceptance of the maintenance recommendation, as illustrated in FIG. 4.

[0090] In a second aspect, there is a method as that in the first aspect, wherein the determining the acceptance of the maintenance recommendation based on the comparisons between the each of the maintenance scheduled time, the maintenance execution time, and the maintenance recommended time comprises, for the comparisons indicative of the maintenance having been executed on or before the scheduled timing and the recommended maintenance timing, determining that the maintenance recommendation was accepted; and for the comparisons indicative of the maintenance having been executed after the recommended maintenance timing and on the scheduled timing, determining that the maintenance recommendation was not accepted as illustrated in FIG. 5.

[0091] In a third aspect, there is a method as that in any of the above aspects, wherein the determining the acceptance of the maintenance recommendation based on the comparisons between the each of the maintenance scheduled time, the maintenance execution time, and the maintenance recommended time comprises, for the comparisons indicative of the maintenance having not been executed before the scheduled timing and after the recommended maintenance timing, determining that the maintenance recommendation was partially accepted as illustrated in FIG. 5.

[0092] In a fourth aspect, there is a method as that in any of the above aspects, wherein the calculating the trust score for the machine operator environment based on the acceptance of the maintenance recommendation comprises incrementing the trust score based on a number of times the maintenance recommendation was accepted over a threshold; and decrementing the trust score based on another number of times the maintenance recommendation was not accepted over another threshold, as illustrated in FIG. 6 (e.g., step 602).

[0093] In a fifth aspect, there is a method as that in any of the above aspects, wherein the calculating the trust score for the machine operator environment based on the acceptance of the maintenance recommendation comprises not incrementing the trust score for instances of when the maintenance recommendation was partially accepted as illustrated in FIG. 6. [0094] In a sixth aspect, there is a method as in any of the above aspects, wherein the calculating the trust score for the machine operator environment based on the acceptance of the maintenance recommendation comprises, for non-execution of the maintenance and an occurrence of failure to the machine, decrementing the trust score; and for non-execution of the maintenance and non-occurrence of failure to the machine, incrementing the trust score as illustrated in FIG. 6 (e.g., step 603).

[0095] In a seventh aspect, there is a method as in any of the above aspects, further comprising modifying an insurance premium for the machine operator environment based on the trust score as illustrated in FIG. 4 (e.g., step 405).

[0096] In an eighth aspect, there is a method as in any of the above aspects, further comprising modifying the trust score from executing a machine learning algorithm configured to classify the machine operator environment into a group from a plurality of groups based on a difference between the maintenance execution time and the maintenance recommended time as well as average loss from execution of the maintenance and failure of the machine from non execution of the maintenance, the machine learning algorithm trained from the timing information received from a plurality of machine operator environments to determine the plurality of groups, the modifying the trust score based on the classification of the machine operator environment into the group as illustrated in FIGS. 8 and 9.

[0097] In a ninth aspect, there is a computer program storing instructions of the method as that in any of the above aspects, and configured to be executed by one or more processors to execute the method steps therein. The computer program and instructions may be stored in a non-transitory computer readable medium.

[0098] In a tenth aspect, there is a system having means for executing any of the method steps described in any of the first through ninth aspects.

[0099] In an eleventh aspect, there is an apparatus having a processor configured to execute any of the method steps described in any of the first through ninth aspects, such as the apparatus illustrated in FIGS. 1 and 2.

[0100] Some portions of the detailed description are presented in terms of algorithms and symbolic representations of operations within a computer. These algorithmic descriptions and symbolic representations are the means used by those skilled in the data processing arts to convey the essence of their innovations to others skilled in the art. An algorithm is a series of defined steps leading to a desired end state or result. In embodiments, the steps carried out require physical manipulations of tangible quantities for achieving a tangible result.

[0101] Unless specifically stated otherwise, as apparent from the discussion, it is appreciated that throughout the description, discussions utilizing terms such as “processing,” “computing,” “calculating,” “determining,” “displaying,” or the like, can include the actions and processes of a computer system or other information processing device that manipulates and transforms data represented as physical (electronic) quantities within the computer system’s registers and memories into other data similarly represented as physical quantities within the computer system’s memories or registers or other information storage, transmission or display devices.

[0102] Embodiments may also relate to an apparatus for performing the operations herein. This apparatus may be specially constructed for the required purposes, or it may include one or more general-purpose computers selectively activated or reconfigured by one or more computer programs. Such computer programs may be stored in a computer readable medium, such as a computer-readable storage medium or a computer-readable signal medium. A computer-readable storage medium may involve tangible mediums such as, but not limited to optical disks, magnetic disks, read-only memories, random access memories, solid state devices and drives, or any other types of tangible or non-transitory media suitable for storing electronic information. A computer readable signal medium may include mediums such as carrier waves. The algorithms and displays presented herein are not inherently related to any particular computer or other apparatus. Computer programs can involve pure software implementations that involve instructions that perform the operations of the desired implementation.

[0103] Various general-purpose systems may be used with programs and modules in accordance with the examples herein, or it may prove convenient to construct a more specialized apparatus to perform desired method steps. In addition, the embodiments are not described with reference to any particular programming language. It will be appreciated that a variety of programming languages may be used to implement the teachings of the embodiments as described herein. The instructions of the programming language(s) may be executed by one or more processing devices, e.g., central processing units (CPUs), processors, or controllers. [0104] As is known in the art, the operations described above can be performed by hardware, software, or some combination of software and hardware. Various aspects of the embodiments may be implemented using circuits and logic devices (hardware), while other aspects may be implemented using instructions stored on a machine-readable medium (software), which if executed by a processor, would cause the processor to perform a method to carry out implementations of the present application. Further, some embodiments of the present application may be performed solely in hardware, whereas other embodiments may be performed solely in software. Moreover, the various functions described can be performed in a single unit, or can be spread across a number of components in any number of ways. When performed by software, the methods may be executed by a processor, such as a general purpose computer, based on instructions stored on a computer-readable medium. If desired, the instructions can be stored on the medium in a compressed and/or encrypted format.

[0105] Moreover, other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the teachings of the present application. Various aspects and/or components of the described embodiments may be used singly or in any combination. It is intended that the specification and embodiments be considered as examples only, with the true scope and spirit of the present application being indicated by the following claims.

Claims

CLAIMS What is claimed is:

1. A method for tracking of maintenance for a machine in a machine operator environment, the method comprising: processing timing information of the machine operator environment from receipt of a maintenance scheduled time indicative of a scheduled timing of the maintenance for the machine and a maintenance execution time indicative of an actual timing of execution of the maintenance for the machine; determining acceptance of a maintenance recommendation based on comparisons between each of the maintenance scheduled time, the maintenance execution time, and a maintenance recommended time indicative of a recommended maintenance timing of the maintenance for the machine; and calculating a trust score for the machine operator environment based on the acceptance of the maintenance recommendation.

2. The method of claim 1, wherein the determining the acceptance of the maintenance recommendation based on the comparisons between the each of the maintenance scheduled time, the maintenance execution time, and the maintenance recommended time comprises: for the comparisons indicative of the maintenance having been executed on or before the scheduled timing and the recommended maintenance timing, determining that the maintenance recommendation was accepted; and for the comparisons indicative of the maintenance having been executed after the recommended maintenance timing and on the scheduled timing, determining that the maintenance recommendation was not accepted.

3. The method of claim 2, wherein the determining the acceptance of the maintenance recommendation based on the comparisons between the each of the maintenance scheduled time, the maintenance execution time, and the maintenance recommended time comprises, for the comparisons indicative of the maintenance having not been executed before the scheduled timing and after the recommended maintenance timing, determining that the maintenance recommendation was partially accepted.

4. The method of claim 1, wherein the calculating the trust score for the machine operator environment based on the acceptance of the maintenance recommendation comprises: incrementing the trust score based on a number of times the maintenance recommendation was accepted over a threshold; and decrementing the trust score based on another number of times the maintenance recommendation was not accepted over another threshold.

5. The method of claim 4, wherein the calculating the trust score for the machine operator environment based on the acceptance of the maintenance recommendation comprises not incrementing the trust score for instances of when the maintenance recommendation was partially accepted.

6. The method of claim 1, wherein the calculating the trust score for the machine operator environment based on the acceptance of the maintenance recommendation comprises: for non-execution of the maintenance and an occurrence of failure to the machine, decrementing the trust score; and for non-execution of the maintenance and non-occurrence of failure to the machine, incrementing the trust score.

7. The method of claim 1, further comprising modifying an insurance premium for the machine operator environment based on the trust score.

8. The method of claim 1, further comprising modifying the trust score from executing a machine learning algorithm configured to classify the machine operator environment into a group from a plurality of groups based on a difference between the maintenance execution time and the maintenance recommended time as well as average loss from execution of the maintenance and failure of the machine from non-execution of the maintenance, the machine learning algorithm trained from the timing information received from a plurality of machine operator environments to determine the plurality of groups, the modifying the trust score based on the classification of the machine operator environment into the group.

9. A non-transitory computer readable medium, storing instructions for tracking of maintenance for a machine in a machine operator environment, the instructions comprising: processing timing information of the machine operator environment from receipt of a maintenance scheduled time indicative of a scheduled timing of the maintenance for the machine and a maintenance execution time indicative of an actual timing of execution of the maintenance for the machine; determining acceptance of a maintenance recommendation based on comparisons between each of the maintenance scheduled time, the maintenance execution time, and a maintenance recommended time indicative of a recommended maintenance timing of the maintenance for the machine; and calculating a trust score for the machine operator environment based on the acceptance of the maintenance recommendation.

10. The non-transitory computer readable medium of claim 9, wherein the determining the acceptance of the maintenance recommendation based on the comparisons between the each of the maintenance scheduled time, the maintenance execution time, and the maintenance recommended time comprises: for the comparisons indicative of the maintenance having been executed on or before the scheduled timing and the recommended maintenance timing, determining that the maintenance recommendation was accepted; and for the comparisons indicative of the maintenance having been executed after the recommended maintenance timing and on the scheduled timing, determining that the maintenance recommendation was not accepted.

11. The non-transitory computer readable medium of claim 10, wherein the determining the acceptance of the maintenance recommendation based on the comparisons between the each of the maintenance scheduled time, the maintenance execution time, and the maintenance recommended time comprises, for the comparisons indicative of the maintenance having not been executed before the scheduled timing and after the recommended maintenance timing, determining that the maintenance recommendation was partially accepted.

12. The non-transitory computer readable medium of claim 9, wherein the calculating the trust score for the machine operator environment based on the acceptance of the maintenance recommendation comprises: incrementing the trust score based on a number of times the maintenance recommendation was accepted over a threshold; and decrementing the trust score based on another number of times the maintenance recommendation was not accepted over another threshold.

13. The non-transitory computer readable medium of claim 12, wherein the calculating the trust score for the machine operator environment based on the acceptance of the maintenance recommendation comprises not incrementing the trust score for instances of when the maintenance recommendation was partially accepted.

14. The non-transitory computer readable medium of claim 9, wherein the calculating the trust score for the machine operator environment based on the acceptance of the maintenance recommendation comprises: for non-execution of the maintenance and an occurrence of failure to the machine, decrementing the trust score; and for non-execution of the maintenance and non-occurrence of failure to the machine, incrementing the trust score.

15. The non-transitory computer readable medium of claim 9, further comprising modifying an insurance premium for the machine operator environment based on the trust score.

16. The non-transitory computer readable medium of claim 9, further comprising modifying the trust score from executing a machine learning algorithm configured to classify the machine operator environment into a group from a plurality of groups based on a difference between the maintenance execution time and the maintenance recommended time as well as average loss from execution of the maintenance and failure of the machine from non-execution of the maintenance, the machine learning algorithm trained from the timing information received from a plurality of machine operator environments to determine the plurality of groups, the modifying the trust score based on the classification of the machine operator environment into the group.

17. An apparatus for tracking of maintenance for a machine in a machine operator environment, the apparatus comprising: a processor, configured to: process timing information of the machine operator environment from receipt of a maintenance scheduled time indicative of a scheduled timing of the maintenance for the machine and a maintenance execution time indicative of an actual timing of execution of the maintenance for the machine; determine acceptance of a maintenance recommendation based on comparisons between each of the maintenance scheduled time, the maintenance execution time, and a maintenance recommended time indicative of a recommended maintenance timing of the maintenance for the machine; and calculate a trust score for the machine operator environment based on the acceptance of the maintenance recommendation.