Industrial equipment operation and maintenance monitoring system
A digital system correlates equipment manufacturing, maintenance, and use history to simulate future performance, addressing the limitations of existing reliability methodologies by providing precise and proactive maintenance recommendations for optimizing equipment behavior and extending lifespan.
Patent Information
- Application Number
- FR2021004418
- Authority / Receiving Office
- FR · FR
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-04-28
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2041-04-28
AI Technical Summary
Existing methodologies for equipment reliability, such as predictive maintenance, physical models, and bench tests, fail to accurately simulate the impact of manufacturing, maintenance, and use on the reliability of industrial equipment, leading to insufficiently precise and proactive predictions and a lack of recommendations for optimizing equipment behavior.
A digital system that generates a behavior model specific to each piece of equipment, correlating manufacturing, maintenance, and use history to simulate future performance, allowing for optimized maintenance decisions and extended lifespan.
The system provides precise, long-term simulations of equipment behavior, enabling optimal maintenance planning, zero-failure operation, and extended equipment life, while reducing uncertainty and optimizing manufacturing and maintenance processes.
Smart Images

Figure 00000080_0000 
Figure 00000081_0000 
Figure 00000082_0000
Abstract
Description
Title of the invention: System for supervising the operation and maintenance of industrial equipment
[0001] The present invention falls within the field of industrial equipment supervision for the purpose of managing their life cycle: - optimize their manufacturing, adapt their manufacturing for a given usage profile; - optimize and adapt their maintenance to a desired usage profile; - determine their optimized maintenance over the long term of their lifespan; - identify uses compatible with no breakdowns in operation until the next maintenance (zero breakdown);
[0002] - identify uses compatible with zero-failure in the long term of their lifetime ; - to instruct and optimize their lifespan extension for a desired usage profile. technical field
[0003] The invention will find preferential, but not limiting, application in the supervision of industrial equipment and installations of medium to high importance, in the sense that the equipment in question is subject to reliability requirements, especially when the reliability requirement is coupled with a safety requirement. This equipment includes, in particular: - those whose unavailability interrupts the production of a service or product of the larger installation to which this equipment belongs; - those whose unavailability induces or aggravates an incidental or accidental situation of the installation.
[0004] The invention thus finds non-limiting applications in the transport segment (road, air, sea, rail), but also in the process industries segment (utilities for "community services", such as the production of electricity, water, fuel), or even in the defense segment.
[0005] In the context of the present invention, the generic term "equipment" thus encompasses a component of an installation, a functional component essential to the proper functioning and / or safety of that installation.
[0006] The equipment considered are therefore components of mobile installations, such as vehicles or craft (for example, submersibles, ships, aircraft or spacecraft), or components of fixed installations, such as structures or infrastructure (for example, power plants, platforms petroleum structures or refineries).
[0007] Because the invention thus allows superior guarantees of reliability in manufacturing, maintenance and operation, the invention meets the strong expectations of industrial players involved in the value chain of this type of equipment, in particular designers, manufacturers, maintainers and operators.
[0008] The reliability of the equipment in question is paramount, since its failure leads to a shutdown of the installation's production. The reliability of such equipment is all the more critical when its reliability is a safety requirement for the installation to which it belongs.
[0009] In this context, the invention takes the following assumptions: the reliability of the equipment depends on the quality of its manufacture, its past maintenance and its past use, as well as the future use that is intended to be made of the equipment. The challenges of equipment reliability thus highlight the importance of controlling the impact of the design and manufacture, maintenance and use of said equipment on its reliability.
[0010] This is why it is so important to answer the following questions: - What are the tasks (gestures, parts and adjustments) for the manufacture of the equipment that optimize the behavior of the equipment in operation, taking into account the intended use? - What are the tasks for equipment maintenance to be carried out that allow the optimum between the cost of maintenance and the potential restored to the equipment for zero breakdowns until the next maintenance (considering the future use that we want to make of the equipment and considering its own manufacturing, maintenance and usage history)? - What is the maximum usage limit for the equipment to be respected during operation until the next maintenance, namely the maximum usage compatible with zero failures until the next maintenance?
[0011] It is also useful to determine at any time the optimal tasks for each of the maintenance operations in the remainder of the equipment's service life, as well as its remaining service life (taking into account said future use and said own history). This allows, in effect: - to individualize the maintenance program for each piece of equipment in the series; - to update and anticipate the supply of spare parts (nature of parts and replacement deadline); - to update and anticipate the programming of industrial maintenance; - to update and anticipate the total cost of future maintenance over the equipment's lifespan.
[0012] It is also useful to determine at any time: - the limit of use of the equipment compatible with zero-failure to be respected on each of the future operating periods in the life of the equipment; - the remaining lifespan of the equipment (taking into account said future use and said own history).
[0013] This allows for realistic planning of the use that can be made of the equipment over the remainder of its life cycle, therefore over the long term.
[0014] In addition, being able to update and anticipate the limit uses of the equipment for the remainder of its life, as well as the total costs of optimal future maintenance, makes it possible to determine at any time the residual value of the equipment with a view to possible resale.
[0015] In summary, the importance of being able to control the impact of manufacturing, maintenance, and use of equipment on equipment reliability is significant, since at a minimum this control: - unlocks access to manufacturing and maintenance optimization; - facilitates interaction with industrial maintenance stakeholders; - unlocks access to mastering equipment reliability (by eliminating uncertainty about reliability in operation); - unlocks access to mastery and optimization of asset management that equipment represents. State of the art
[0016] To date, the existing methodologies best suited to addressing equipment reliability are: - predictive maintenance; - physical models; - qualification programs on test benches; - feedback (commonly referred to as "REX"). However, these methodologies lack the simulation capabilities required to fully understand the impact of manufacturing, maintenance, and equipment use on its reliability. In this respect, they only partially meet the needs of manufacturers. Regarding Predictive Maintenance, this technology observes and analyzes equipment operation, waiting to detect any degradation. Once degradation is detected, it establishes an initial trend in the evolution of the degradation to provide a preliminary prediction of the timeframe for a future failure. This trend and the predicted timeframe are then refined iteratively with subsequent observations.
[0017] In doing so, Predictive Maintenance involves several methodological biases: - it seeks to deduce the future behavior of the equipment from the observed past behavior, thus seeking to deduce the future consequence of aging, independently of the past and future causes of aging; - it assumes that future use will be identical to past use - which is almost always false; - it uses operational data alone, independently of functional expertise.
[0018] Therefore, Predictive Maintenance is: - approximate and slow in its initial predictions, requiring an observation period of approximately fifteen days of operation following maintenance; - slow to display an accurate prediction; - myopic: its prediction horizon is limited to a medium-term duration, at most on the order of six months.
[0019] As a further consequence of its methodological biases, Predictive Maintenance is insufficiently proactive. Indeed, it only reacts to the occurrence of degradation during operation and therefore does little to anticipate the occurrence of degradation. Moreover, it requires the equipment to be operational in order to develop its predictions: this is a significant operational drawback. More generally, Predictive Maintenance lacks access to any simulation, in particular the simulation of equipment behavior based on its manufacturing or maintenance history and the anticipated use of that equipment.
[0020] Predictive Maintenance is therefore limited to providing only a prediction of failure time, which is insufficiently precise and proactive. It lacks the capacity to provide recommendations in terms of manufacturing, maintenance, or use to optimize equipment behavior, particularly with a view to postponing any failure time or extending service life.
[0021] The various Predictive Maintenance solutions have developed with the rise of computer resources and the increasing value of data. However, they only exploit operational data, are affected by the same methodological biases, consequently suffer from the same limitations, and are only slightly differentiated from one another. Regarding physical models, the complete and individualized modeling of a given piece of equipment using physical models is not realistic from either a technical or, especially, an economic standpoint. These models are very expensive and cover too limited a range of physical phenomena. Furthermore, they are approximate.
[0022] Indeed, a physical model is only able to characterize the behavior of an elementary part of a component of equipment subjected to a physical phenomenon. This modeling is also an approximation of reality: it contains a degree of inaccuracy. Associated with fundamental research, the development of a physical model also requires considerable resources in terms of time and cost. Furthermore, to characterize the aging of each of the elementary parts of a piece of equipment, before arriving at a global model of the equipment, it would also be necessary to generate a multitude of physical models, on the order of several hundred.
[0023] Modeling the behavior of a typical piece of equipment in a series using physical models therefore requires considerable resources in terms of cost and time. Furthermore, this overall modeling based on these physical models is difficult to individualize to the case of each piece of equipment, particularly its own complete history. Regarding bench tests, carried out during the design phase of the reference equipment in a given series, especially during its qualification, these tests allow for a theoretical approximation of the behavior of the reference equipment throughout its lifespan, under one or more constraints simulating use and aging. These tests provide trends and orders of magnitude: the indication is therefore only approximate.
[0024] Decoupled from the specificity of a given piece of equipment in the series taken in its individuality (that is to say not taking into account the manufacturing, maintenance and use history of a given piece of equipment which these tests cannot in fact anticipate), these tests cannot really characterize the future behavior of this equipment, in response to the future use which will be made of it, much less determine the maintenance parameters to be carried out which would optimize the behavior of this equipment in operation.
[0025] Indeed, bench tests consist of characterizing the behavior of the reference equipment in the series under consideration (for example, a steady-state operating point or a response to a transient) when subjected to one or more given usage and aging profiles: the representative sampled equipment from the series is then subjected to accelerated aging, considered representative of actual aging. The characterization of the behavior remains approximate, because aging during the test is accelerated and the usage considered during the test is distinct from the actual uses that will be made of each piece of equipment in the series during its respective lifespan.
[0026] The behavior of the reference equipment of the series and in particular its lifespan are thus only approximated. Regarding feedback, commonly referred to as "REX," it complements the trends and orders of magnitude approximated by bench tests. REX cannot Nor does it take into account the specific characteristics of a given piece of equipment within the series, particularly its complete history, nor therefore characterize the future behavior of the equipment in response to a given future use, much less determine the maintenance parameters that will optimize its operating performance. Furthermore, the data from the feedback process is insufficient due to its unsystematic and insufficiently regular use. Indeed, REX (Return on Experience) involves comparing an average behavior, such as lifespan, to average usage. This approach does not allow for obtaining precise data. Furthermore, the REX is not systematically used. When the REX is used, it is used irregularly and the frequency of REX updates is too low, associated with a periodicity of around one to two years.
[0027] The invention aims to overcome the drawbacks of the prior art, by proposing a digital system for supervising the operation and maintenance of at least one piece of industrial equipment within an installation, capable of generating and using a behavior model specific to the equipment in the series.
[0028] This generated behavioral model is relevant: - it correlates the causes and consequences of equipment aging; - it is specific to the physical phenomena of which each piece of equipment in the series is the site; - it can be customized for each piece of equipment in the series; - it has the accuracy of empirical models that are learned from observed data ("data-driven" models).
[0029] The generated behavioral model is powerful: it allows for the simulation of the behavior of a piece of equipment from the series, for a given maintenance decision and a given predicted use of the equipment, taking into account the manufacturing, maintenance, and usage history of the equipment. The model is all the more powerful because it allows this type of simulation over the long term of the equipment's lifespan.
[0030] For a given piece of equipment in the series considered, taking into account the intended intended use of the equipment and its manufacturing, maintenance and usage history, the invention thus makes it possible to obtain overall value propositions, including: - For a given equipment maintenance stage, the system determines the optimal maintenance decision. This decision is the one that strikes the best balance between maintenance constraints and the equipment's potential for zero downtime until the next maintenance and for the desired projected usage. This optimal decision is also tailored to the historical data. manufacturing, maintenance and use of the equipment; - for a period of equipment operation during its lifetime, the system makes it possible to determine the limit use, that is to say the maximum possible use compatible with zero failure until the next maintenance; - the system allows, during the operation of the installation, to update the limit use of the equipment, taking into account the actual use of the equipment since its last maintenance and taking into account the desired forecast use until the next maintenance; - the system also allows to determine at any time the program of optimal future maintenance of the equipment over the remainder of its life as well as the limit uses of the equipment to be respected for each period of operation over the remainder of its life; - the system determines the end-of-life date of the equipment at any time; - Finally, the system makes it possible to determine the manufacturing tasks (gestures, parts, settings) that optimize the life of the reference equipment in the series, for a given forecast usage scenario, in particular a scenario dependent on the usage segment crossed with a geographical usage segment.
[0031] To this end, the invention operates on the scale of a series of equipment. For a given series, the invention first determines a correlation between the causes and consequences of aging of the equipment in that series. This correlation links the manufacturing and maintenance history and the usage history of a given piece of equipment in the series to the condition of that equipment as determined by these histories. The correlation determined is specific to the physical phenomena affecting each piece of equipment in the series. Secondly, the invention models this correlation in the form of a virtual model learned from the manufacturing, maintenance, usage, and condition data of the equipment in that series.Thirdly, the invention uses this model as a simulator individualized for each piece of equipment in the series: the model makes it possible to simulate the state of the equipment at a given moment for a maintenance decision and for the use that will follow up to that moment, taking into account the manufacturing and maintenance history and the usage history specific to the equipment and preceding the maintenance in question. This model provides access to the value propositions of the invention.
[0032] Furthermore, by modeling the behavior of the standard equipment in the series in this way, the invention correlates, at the series level, the observed behavior of the equipment with the manufacturing, maintenance, and use performed on each of said pieces of equipment: the invention is therefore a form of exploiting feedback. This feedback The superior character of experience derives from the relevance, from a functional point of view, of the modeled correlation. The invention also allows for the continuous or near-continuous exploitation of systematized feedback, by restarting the model's learning from updated manufacturing, maintenance, use, and condition data of the equipment in the series under consideration, according to a periodicity determined as most suitable by the designer, manufacturer, maintenance provider, and operators.
[0033] To this end, the invention aims at a digital system for supervising the operation and maintenance of at least one piece of industrial equipment within an installation, executed by at least one computer terminal. Advantageously, the said system includes at least the following steps: - an installation is carried out within an installation of at least one piece of industrial equipment from a manufacturing process and representative of a series, then said equipment is operated at least within a period until a maintenance step is carried out; - at least one planned maintenance is defined after the said maintenance to be carried out, following at least one scenario with planned conditions of use of said equipment over a planned operating period. Such a system is characterized in that the manufacture, installation, operation and maintenance of said equipment entail at least: - a manufacturing and maintenance history including: manufacturing tasks for said at least one piece of equipment up to said installation; possibly tasks for at least one prior maintenance of said equipment, - a history of use of said equipment over said period between installation and said maintenance to be carried out, said history of use including conditions of use of said equipment during said period; - a state history of said equipment, said state history including hardware indicators of said equipment; in that - by means of a technical analysis of said equipment, at least one correlation is determined between, on the one hand, at least one of said tasks and / or at least one of said conditions of use and, on the other hand, at least one of the material indicators of said condition, said correlation establishing at least one link between causes of aging and consequences of aging of the equipment; in the correlation: - Tasks are characterized by those identified as critical, and / or conditions of use by those to which the equipment is sensitive and exposed during operation. whether in operation or stopped, the tasks and conditions considered impacting the state of said equipment; - the physical condition of the equipment is characterized by indicators identified as being representative of the condition of said equipment; Next, in the correlation, we determine: - measured physical quantities or functions of measured physical quantities characterizing the conditions of use to which said equipment is sensitive and exposed in operation or when stopped; - measured physical quantities or functions of measured physical quantities characterizing the material state of the equipment at a given instant; Then - for the other equipment in the said series, data associated with these tasks, and data associated with physical quantities or functions of physical quantities relating to these conditions of use and these material indicators, as identified in the said correlation, are retrieved and extracted in order to obtain a dataset; - we train at least one virtual model, based on the dataset. The system is further characterized by the fact that During the maintenance to be carried out on said equipment, the following values are submitted to the model: - on the one hand, at least one of the tasks of the manufacturing and maintenance history and at least one of the conditions of use of the usage history, - on the other hand, at least one of the maintenance tasks to be carried out and the said predictive conditions of use of the scenario; said model generating a predicted state of said equipment after said maintenance to be carried out, said predicted state being compared to a minimum state identified as required for the operation of said equipment. Presentation of the figures
[0034] Other features and advantages of the invention will become apparent from the following detailed description of non-limiting embodiments of the invention, with reference to the accompanying figures, in which:
[0035] [fig. 1] schematically represents a view of a system architecture implemented in the supervision of equipment, showing in particular, on the eve of maintenance to be carried out on the equipment, the submission to the learned virtual model, on the one hand, of the manufacturing and maintenance history and the usage history of said equipment, and on the other hand, of the maintenance tasks to be carried out and the conditions forecasts of use of the scenario, the said virtual model generating a forecast state of said equipment at the due date of the next forecast maintenance;
[0036] [fig.2] schematically represents a view of a detail of the architecture, showing in particular, on the one hand a manufacturing and maintenance history including the decision to manufacture the equipment up to its installation, as well as the decisions for prior maintenance of the equipment, and on the other hand the decisions for maintenance to be carried out and those for planned maintenance, highlighting in particular the tasks of each of said decisions and their timing;
[0037] [fig.3] schematically represents a view of a detail of the architecture, showing in particular, on the one hand the history of use of the equipment between installation and maintenance to be carried out, and on the other hand the forecast scenario of use, highlighting that said history of use and said scenario include conditions of use of said equipment associated with their respective period;
[0038] [fig.4] schematically represents a view of a detail of the architecture, showing in particular, on the one hand the history of the successive states of the equipment, and on the other hand the successive forecast states of the equipment, highlighting in particular the material indicators of the states of the equipment associated with their respective period;
[0039] [fig.5] schematically represents a view of a detail of the architecture of said supervisory system, showing in particular the process of retrieving and extracting data from a piece of equipment in the series and as identified in the correlation to be learned;
[0040] [fig.6] schematically represents a view of another architectural detail, showing in particular the process of retrieving and extracting data for several pieces of equipment in the series, in order to obtain a dataset and learn the virtual model;
[0041] [fig.7] schematically represents a view of an architecture of the system of su pervision, showing in particular said virtual model generating at a given moment a predicted state of said equipment, and its two hardware indicators, with a comparison with a minimum state identified as required for the operation of said equipment;
[0042] [fig.8] schematically represents a view of an example of a limit calculation for a predicted condition of a usage scenario since a maintenance to be performed, limit for which the predicted state of the equipment is equivalent to the minimum state, at the time of the next predicted maintenance;
[0043] [fig.9] schematically represents another view of said example of another calculation of another limit for another predicted condition of a usage scenario since maintenance to be performed, another limit for which the predicted state of the equipment is equivalent to the minimum state, at the time of the predicted maintenance- following section;
[0044] [fig.10] schematically represents a view of said example with a scale of values of certain anticipated conditions of use, in the form of polygons, indicating in particular the target values and maximum limits for said anticipated conditions associated with said scenario;
[0045] [fig.l 1] schematically represents a view of said example with a network of curves in abacus, highlighting in particular, from the variation of two of the forecast conditions, the maximum limit of a third of the forecast conditions;
[0046] [fig. 12] schematically represents a view similar to figure 1 of a long-term architecture, showing in particular a first step of recurrent implementation of said supervisory system to predict forecast states, from maintenance to be carried out to predictive maintenance;
[0047] [fig. 13] schematically represents a view similar to figure 12 for the long term, showing in particular a second stage of implementation in a recurrent manner of said supervisory system to predict forecast states, from said predictive maintenance until a subsequent predictive maintenance;
[0048] [fig. 14] schematically represents a simplified view of an example of long-term prediction, showing in particular curves representing the evolution of the state of the equipment as a function of time, during two consecutive forecast periods;
[0049] [fig. 15] schematically represents a simplified view of an example of long-term prediction, showing in particular curves representing the evolution of the equipment's condition over time, during several consecutive forecast periods; and
[0050] [fig. 16] schematically represents a simplified view of a supervisory system architecture, showing in particular said virtual model determining the time of a failure deadline of said equipment, for an insufficient maintenance decision. DETAILED DESCRIPTION
[0051] The present invention relates to a system 1 for supervising the operation and maintenance of at least one piece of equipment 2 within an industrial installation 3.
[0052] Such a supervisory system 1, hereinafter referred to as "system 1", is intended to be digital. In other words, it consists of at least one software program, intended to be executed by at least one computer terminal. Typically, such a computer terminal can be of any type, including a computer server or a computer. Furthermore, this computer terminal allows the through suitable storage means: recording, reading, modifying and generating data in digital form. This computer terminal is also designed to be accessible via a suitable communication network, locally or remotely.
[0053] Furthermore, such a system 1 comprises successive steps, described below in a non-exhaustive manner. According to the invention, the system 1 can therefore be considered a method. Some steps are carried out in real terms, in particular by a person interacting with said computer terminal, in particular via a virtual interface, or specifically when an operator interacts with said equipment 2. Other steps are performed virtually, particularly when data processing is carried out by the computer terminal. This system 1 therefore involves digital elements, resembling virtual technical means, advantageously implemented within the framework of the present invention.
[0054] Said system 1 ensures the supervision of a piece of equipment 2 within an installation 3, of several pieces of equipment 2 within the same installation 3, or of several pieces of equipment 2 within several installations 3.
[0055] As mentioned previously, said equipment 2 is characterized as essential for the operation of said installation 3. Therefore, the unavailability of equipment 2 is likely to interrupt the production of the installation 3, to induce or aggravate an incidental or accidental situation of the installation 3. Said equipment 2 can be of any type, for example a thermal engine, an electric motor, an alternator, a pump, a solenoid valve.
[0056] Said equipment 2 is therefore an integral part of the installation 3 and is necessary for its proper functioning. Such an installation 3 can be mobile, like a machine or vehicle (for example a submersible, ship, aircraft or spacecraft), or fixed, like a structure or infrastructure (for example a power generation plant, an oil platform or a refinery).
[0057] Therefore, system 1 includes, as an initial condition, the installation within an industrial installation 3 of at least one piece of equipment 2 resulting from a manufacturing process 4 and representative of a series. In other words, the equipment 2 was integrated at some point in the past, within the installation 3. Furthermore, the manufacturing step 4 of equipment 2 includes the assembly of several components until the completed equipment 2 is obtained, followed by the installation of the equipment 2. This manufacturing step 4 is preceded by a design step 40 of the equipment 2.
[0058] It should be noted that equipment 2 is representative of a series, which includes All equipment 2 is manufactured identically from the same design plan. A fraction of the equipment 2 in the series may no longer be in use, another fraction is in use, and another fraction, already manufactured, may remain to be installed and operated.
[0059] Following its installation in the facility 3, said equipment 2 is operated for at least a period 5 until a maintenance step 6 is performed. Therefore, said period 5 corresponds to a period of operation of the equipment between installation and the maintenance 6 to be performed, or to an alternation of periods of operation and maintenance steps since installation.
[0060] It should be noted that the maintenance step 6 to be carried out is a concrete operation, requiring the intervention of at least one operator on the geographical site of the installation 3 and carried out directly on said equipment 2.
[0061] In this context and as mentioned previously, the supervisory system 1 plans to manage the future lifecycle of said equipment 2, namely: - define the nature of the tasks for future maintenance of equipment 2 in order to optimize them and adapt them to the desired usage profile of said equipment 2, and this in the long term and for the life of equipment 2; - identify the limit uses compatible with zero-failure in operation until the next maintenance for each planned operating period of equipment 2, and this in the long term and for the life of equipment 2; - update in real time, during operation, the limit usage of equipment 2, compatible with zero-failure; - instruct the extension of the life of equipment 2, by defining the nature of the tasks of the last maintenance of equipment 2 optimized and adapted to the desired usage profile of equipment 2.
[0062] At the scale of the equipment in the series, the supervisory system 1 also plans to optimize the manufacture of the representative type of equipment in the series to adapt it to a given usage profile, namely to determine the manufacturing parameters 4 which optimize the life cycle of the equipment 2.
[0063] Therefore, at least one planned maintenance 7 is defined after the said maintenance 6 to be carried out. The upcoming period between the aforementioned maintenance 6 to be performed and the aforementioned planned maintenance 7 corresponds to at least one planned operating period 70 during which the equipment 2 is operated under at least one scenario 8 characterized by planned operating conditions 110, referred to as "planned conditions 110". The upcoming period may also consist of alternating planned operating periods and maintenance phases. The coming period may extend until the end of the product's lifespan. the equipment.
[0064] Figure 1 shows in particular the design 40 and the manufacture 4 of the equipment 2, the period 5 (between the installation of said equipment 2 in the installation 3 until the maintenance 6 to be carried out and including any prior maintenance 10), the planned period 70 until a planned maintenance 7.
[0065] That being said, the manufacturing steps 4 and the installation of said equipment 2, and then the steps of operation and maintenance of said equipment 2 during period 5 generate several types of data which the supervisory system 1 takes into consideration.
[0066] The manufacture, installation, operation and maintenance of said equipment 2 result in at least one manufacturing and maintenance history 9 (or "history 9") comprising: - 90 manufacturing tasks of said equipment 2 up to said installation; - possibly tasks 90 of at least one prior maintenance 10 of said equipment 2.
[0067] In other words, the manufacturing and installation steps include at least 90 manufacturing and installation tasks. Furthermore, period 5 may include at least one prior maintenance 10 of said equipment 2, which includes 90 maintenance tasks. All the manufacturing tasks 90 of said at least one piece of equipment 2 up to said installation, as well as any tasks 90 of at least one prior maintenance 10 of said equipment 2, generate a manufacturing and maintenance history 9.
[0068] It should be noted that the manufacturing tasks 90 and the maintenance tasks 90 of each of the maintenance 10 are defined in a non-limiting way by actions performed by an operator, by parts assembled during manufacturing or disassembled at a maintenance stage to be replaced, as well as by adjustments of said equipment 2.
[0069] As can be seen in Figure 2, system 1 includes in particular: - the manufacturing and maintenance history 9, that is to say the history of manufacturing and maintenance decisions executed under manufacturing 4 and implementation (of rank 0) and under prior maintenance 10 (from rank 1 to rank j-1); - a maintenance decision under maintenance 6 to be carried out (of rank j); - a maintenance decision under the planned maintenance 7 (of rank j+1). Each manufacturing or maintenance decision of rank k is characterized by its relation to the different possible tasks 90 (completion or not of said task 90, characterization of said task 90 when performed time counter 91 characterizing the duration since said task 90 has been performed on equipment 2).
[0070] Furthermore, these steps also generate a usage history 11 (or "history 11") of the equipment 2 over the period 5 between said installation and said maintenance 6 to be performed. This usage history 11 includes conditions 110 of the use of said equipment 2 during said period 5 (also referred to as "conditions 110"). Indeed, the period 5 between the installation of equipment 2 and the maintenance 6 to be carried out includes at least one period of operation of equipment 2 in the context of use.
[0071] This use is associated with the way in which the equipment was used, namely with all the conditions 110 of use of said equipment 2 during said period.
[0072] For example, in the case of equipment 2 corresponding to the engine of a truck-type installation 3, these conditions 110 of use may be the load transported, the mileage travelled, the speed, the ambient air temperature, the average slope of the roads used.
[0073] This use thus generates the usage history 11 of said equipment over the period 5 between installation and said maintenance 6 to be carried out. As can be seen in Figure 3 illustrating the case of a pump-type piece of equipment 2 (for which one of the conditions 110 is temperature), the system 1 includes in particular: - the history 11 of use of equipment 2 during period 5 (from the implementation of rank 0 to the maintenance 6 to be carried out of rank j); - Scenario 8 of equipment 2 use as planned over the forecast period 70. The usage history 11 can be broken down into the set of usage sub-histories 111 of equipment 2. The usage sub-history 111 of rank k represents the fraction of the usage history 11 of equipment 2 during the operating period between the two successive prior maintenance 10s of rank k and rank k+1. In particular, Figure 3 highlights: - the sub-history 111 of rank 0 usage (between the rank 0 implementation and the prior rank 1 maintenance 10); - the sub-history 111 of usage rank j-1 (between the last prior maintenance 10 of rank j-1 and the maintenance 6 to be carried out of rank j).
[0074] History 11 includes all sub-histories 111 of use from rank 0 to rank j-1. The k-rank usage subhistory 111 represents the evolution of each of the 110 usage conditions between two previous 10 maintenances of rank k and rank k+1, such as the evolution of temperature over time (as seen in Figure 3). In sub-history 111 of rank k equipment 2 usage, one can distinguish the partial sub-history 1110 of equipment 2 usage of rank (k,t), that is to say the evolution of each of the conditions 110 of use between the prior maintenance 10 of rank k and the instant (t), for an instant (t) between the maintenance of rank k and the maintenance of rank k+1. Scenario 8 represents the predicted evolution of each of the predicted conditions 110 of use, within the framework of the predicted use of equipment 2 between the maintenance 6 to be carried out (rank j) and the predicted maintenance 7 (rank j+1), as seen in figure 3. In scenario 8, we can distinguish a partial scenario 80.The partial scenario 80 of rank (j,t) represents the predicted evolution of each of the conditions 110 of predicted use of equipment 2 between the maintenance 6 to be carried out of rank j and the instant (t), for an instant (t) between the maintenance 6 to be carried out of rank j and the predicted maintenance 7 of rank j+1. .
[0075] Moreover, said steps also induce a history 12 of the state of said equipment 2 (or "history 12") which includes material indicators 120 of said equipment 2 (or "indicators 120"). Indeed, the use of equipment 2 induces aging that impacts the state 13 of equipment 2. The state 13 of equipment 2 at a given moment, characterized by the hardware indicators 120, reflects the physical integrity of equipment 2, upon which its ability to function properly depends. The set of states 13 of equipment 2 (i.e., the set of hardware indicators 120) that were successively in effect during period 5 constitutes the state history 12 of equipment 2.
[0076] As can be seen in Figure 4 illustrating the case of a pump-type piece of equipment 2 (for which one of the material indicators 120 is the flow rate (Q)), the system 1 includes, in particular, the state history 12 of the equipment 2 during the Period 5. The state history 12 can be decomposed as the set of state sub-histories 121. The state sub-history 121 of rank k for equipment 2 represents the fraction of the state history 12 during the operating period between the two successive prior maintenance 10s of rank k and rank k+1. In particular, Figure 4 highlights: - the sub-history 121 of rank 0 state (between the implementation at rank 0 and the prior maintenance 10 of rank 1); - the sub-history 121 of state of rank j-1 (between the last maintenance 10 prior to rank j-1 and the maintenance 6 to be carried out at rank j).
[0077] History 12 comprises all the state sub-histories 121 from rank 0 to rank j-1. As shown in Figure 4, the state sub-history 121 of rank k for equipment 2 represents the evolution of the successive states 13 of equipment 2, between the two prior maintenance 10s of rank k and rank k+1. The state sub-history 121 of rank k for equipment 2 represents the evolution of each of the indicators 120 material (for example, the evolution of the flow rate over time) between the two prior maintenance 10s of rank k and rank k+1. As can be seen in Figure 4, we can distinguish, in the sub-history 121 of the rank k state of equipment 2, the partial state sub-history 1210. The partial state sub-history 1210 of rank (k,t) represents the evolution of each of the material indicators 120 between the prior maintenance 10 of rank k and time (t) (for a time (t) between the maintenance of rank k and the maintenance of rank k+1).
[0078] Thus, the historical data 9, 11, 12 extend over time during period 5. They include respectively the manufacturing and maintenance tasks 90, the operating conditions 110 and the material indicators 120.
[0079] The aforementioned historical elements 9, 11, 12 represent computer field names, within which values measured from said equipment 2 were recorded successively over time.
[0080] Advantageously, initially, at least one correlation 14 is determined between, on the one hand, at least one of said manufacturing and maintenance tasks 90, and / or at least one of said conditions 110 of use and, on the other hand, at least one of the material indicators 120 of said state 13. Said correlation 14 establishes at least one link between causes of aging and consequences of aging of the equipment 2. In other words, the invention chooses to approach the behavior of equipment 2 from the perspective of the causes and consequences of aging.
[0081] For this purpose, the invention chooses on the one hand to characterize the causes of the aging of the equipment 2 by the history 9 of manufacture and maintenance of the equipment 2, as well as by the history 11 of use of the equipment 2. The invention also chooses to characterize the consequences of the aging of equipment 2 by the state 13 of equipment 2 induced by said histories 9 and 11.
[0082] The invention therefore takes the approach of correlating the state 13 of equipment 2 with the manufacturing and maintenance history 9 and the usage history 11 of equipment 2.
[0083] Thus, the invention provides for correlating between them the manufacturing and maintenance history 9 (hereinafter referred to as "first term" of said correlation 14) and the usage history 11 (hereinafter referred to as "second term" of said correlation 14) of a given piece of equipment 2 from the series to the state 13 of this equipment 2 (hereinafter referred to as "third term" of said correlation 14) induced by these histories 9, 11 - said three terms of said correlation 14 constituting a triplet 140.
[0084] For example, consider the case of a pump-type device 2 that is sensitive to the temperature of the conveyed fluid. The invention correlates, on the one hand, the pump's usage history (characterized by the fluid temperature and / or pressure history) suction and / or speed of the pump) as well as the main pump manufacturing options (such as the type of impeller fitted) and the pump maintenance history (such as the history of impeller replacements at the various previous maintenance 10) on the other hand the condition of the pump (characterized by its flow rate and / or discharge pressure).
[0085] Further on, in correlation 14, only the relevant parameters likely to influence the aging and operation of equipment 2 are considered, the other parameters not being appropriate to select. To do this, we carry out a technical analysis of equipment 2 to identify the critical manufacturing and maintenance tasks 90 as well as the conditions 110 of use which impact the state 13 of equipment 2.
[0086] To do this, by means of the technical analysis of the equipment 2, the manufacturing and maintenance tasks 90 are characterized by those identified as critical, as defined below.
[0087] More precisely, the invention chooses to characterize the tasks 90 of the manufacturing and maintenance history 9 of the equipment 2 in at least one of the following ways: - for the manufacturing 4 of the equipment 2, the invention characterizes the actions performed, the parts assembled, or the settings adopted during manufacturing. The manufacturing actions are characterized by the protocol option for each manufacturing action, if there are several possible protocol options for said action for the manufacturing 4 of the equipment 2 in the series. The parts assembled during manufacturing are characterized by the part assembly option, if there are several possible options for said part for the manufacturing 4 of the equipment 2 in the series (for example: if the manufacturing of the series of pumps provides for two possible types of pump bearing).The settings adopted during manufacturing are characterized by the value of each setting, if there are several possible values of said setting for the manufacture 4 of the equipment 2 in the series (for example: the tightening torque of the pump packing gland); . - For a maintenance step of equipment 2, the invention characterizes the actions performed, the parts installed, or the settings adopted at each maintenance step. Among the maintenance actions, part replacement is distinguished from other maintenance tasks (for example: tightening the connections of an electrical terminal block). Part replacement during the maintenance considered is characterized by the part option, if there are several possible options for said part for the maintenance of equipment 2 in the series (for example: if the pump series provides two possible types of pump bearing). The settings adopted during the maintenance considered are characterized by the value of each setting, if there are several possible values of said setting for the maintenance of equipment 2 in the series (for example, the tightening torque of the pump stuffing box).
[0088] Furthermore, and according to one embodiment, in correlation 14, the manufacturing and maintenance history 9 is reduced to a history of critical tasks 90 in the form of at least a list of successive values.
[0089] In other words, and by means of the technical analysis of the equipment 2, the invention chooses to characterize the manufacturing and maintenance history 9 of the equipment 2 by considering only the critical tasks 90, namely the actions, parts and adjustments identified as determining the behavior of the equipment 2 in operation (i.e. as having an impact on aging and thus on the condition 13 of the equipment 2).
[0090] Furthermore, the manufacturing and maintenance history 9 of equipment 2 (up to and including the last prior maintenance 10) is characterized by listing, for each critical task 90, the values that were successively adopted during manufacturing 4 and during the various successive prior maintenance 10s in the life of equipment 2, up to and including the last prior maintenance 10 (the one preceding the maintenance 6 to be performed). The manufacturing and maintenance history 9 is thus characterized by a list of values. For example, consider the case of a pump-type piece of equipment 2 with a bearing, for which the installation or replacement, as well as the type of pump bearing, corresponds to a critical manufacturing and maintenance task 90 for the pump. In this example, the last prior maintenance 10 (the one preceding the maintenance 6 to be performed) is considered to be the seventh (of rank k=7).The invention characterizes the history of this critical manufacturing and maintenance task 90 by the list [A,0,0,B,0,0,B,0], to represent the installation of the type A bearing during manufacturing 4, the replacement of the bearing with a new type B bearing during the third maintenance, the replacement of the bearing with a new type B bearing during the sixth maintenance, as well as the fact that no maintenance action was performed on the bearing during the other prior maintenance 10. The invention therefore chooses to characterize the manufacturing and maintenance history 9 of equipment 2 by a matrix consisting of lists, each list being associated with a critical task 90 and recording the successive values characterizing this critical task 90 throughout manufacturing 4 and then the various maintenance of equipment 2 up to and including the last prior maintenance 10.For example, consider the aforementioned case of a pump-type piece of equipment 2 where the manufacturing and maintenance history 9 can be reduced to the history of the two critical tasks 90 "installation or replacement of the pump bearing" and "installation or replacement of the pump impeller", and with the respective history lists after the seventh maintenance being [C,0,A,0,0,B,0,C] and [A,0,0,B,0,0,B,0]. The manufacturing and maintenance history 9 of the pump under consideration up to and including the seventh maintenance then corresponds to the matrix. [ [C,0,A,0,0,B,0,C] ; [A,0,0,B,0,0,B,0]].
[0091] In addition, the conditions 110 of use are characterized by those to which the equipment 2 is sensitive and exposed in operation or when stopped.
[0092] Indeed, by means of the technical analysis of the equipment 2, the invention reduces the conditions 110 to the ambient conditions and / or operating conditions (AC / CF) to which the equipment 2 is sensitive and exposed in operation or when stopped. Ambient conditions (AC) are understood to mean conditions of the external environment of equipment 2 and to which equipment 2 is sensitive and exposed in operation or when stopped (such as ambient air temperature, humidity, irradiation rate). The operating conditions (OC) are understood to mean: - the internal conditions of equipment 2 to which equipment 2 is sensitive and exposed in operation or when stopped (such as the temperature of the conveyed fluid, the level of vibration in the case of a pump); and / or - the parameters representing the power deployed by the equipment 2 in operation and which impact its operating point (such as the load transported, the speed in the case of a truck engine); and / or - the other parameters necessary to characterize the amount of work to which the equipment 2 was subjected under the aforementioned ambient and operating conditions (AC / CF) (such as the total mileage travelled in the case of a truck, or generally the time of use of the equipment 2).
[0093] For example, consider the case of a pump-type equipment 2 whose impeller is made of thermoplastic material and is therefore sensitive to and exposed to the temperature of the conveyed fluid. The temperature of the conveyed fluid is then all the more important to take into account as an operating condition than in the case of a pump whose impeller is made of metal.
[0094] In the sense of the invention, scenario 8 and the anticipated conditions 110 of use are characterized by the same ambient and operating conditions (AC / CF) to which the equipment 2 is sensitive and exposed in operation or when stopped.
[0095] Furthermore, the material state 13 of the equipment 2 is characterized by the material indicators 120 identified as being representative of this material state 13 of said equipment 2.
[0096] To this end, and by means of the technical analysis of the equipment 2, the invention chooses to characterize the state of the equipment 2 at the instant by the necessary and sufficient set of performance, vibration behavior and other indicators 120 materials (for example, and not limited to, material indicators usually measured by non-destructive testing techniques), deemed representative) of the condition 13 of the equipment 2.
[0097] To characterize the correlation 14 and in particular the history of the conditions 110, in the most specific way possible to the equipment 2 considered, we determine the measured physical quantities or the functions of the measured physical quantities characterizing the conditions 110 to which said equipment 2 is sensitive and exposed in operation or at rest or characterizing as best as possible the history of these conditions 110 of use (i.e. the history 11 of use). These functions of physical quantities comprise several mathematical or algorithmic functions. In other words, each condition 110 is associated with a measured physical quantity that characterizes that condition. Thus, where relevant, the invention characterizes the history of the condition by the history of this physical quantity. For example, if equipment 2 is sensitive to and exposed to ambient air temperature, or to the temperature of the fluid conveyed in the case of a pump, then the corresponding temperature history is considered. Furthermore, the invention also chooses to characterize the history of a condition 110 by the history of a function of the physical quantity representing that condition 110, when this characterization is more relevant than the history of that physical quantity itself. For example, if equipment 2 is sensitive and exposed to ambient air temperature or to the temperature of the conveyed fluid (as in the case of a pump), then the temperature history can be assimilated to the time integral of the temperature over the period between the installation of equipment 2 in installation 3 until the moment. Further, when the invention chooses to characterize the history of a condition 110 by the history of a function of the physical quantity representative of said condition, these functions may include, without limitation, a calculation of the time of presence of the measured physical quantity in at least one range of values.
[0098] Indeed, the invention takes the position that the accumulation of abnormal transients impacts the aging and therefore the behavior of the equipment 2. Thus, when relevant, the invention characterizes the history of a condition 110 by counting the times of presence of this condition respectively in the normal operating domain, in the domain close to destruction, or even in the domain intermediate to the two previous ones.
[0099] For example, consider the case of equipment 2 of the pump type that is sensitive and exposed to the temperature of the conveyed fluid. The history of the temperature by the time integral of the temperature over the period between the installation of equipment 2 in installation 3 and the current time, distinguishing the component of said integral within the normal operating range, the component within the near-destruction range, and the component within the intermediate range between the two preceding ranges. When the invention chooses to characterize the history of a condition 110 by the history of a function of the physical quantity representing said condition, these functions may also include, but are not limited to, a representative calculation of at least one fluctuation of the measured physical quantities and / or a counting of said at least one fluctuation. This is particularly the case when equipment 2 is sensitive and exposed to variations of a condition 110. Without limitation, such functions can then correspond to the gradient function or the cycle count of the condition considered.
[0100] As a first example, we consider the case of a pump-type piece of equipment 2 that is sensitive to and exposed to the temperature of the fluid conveyed, and in particular to sudden variations in said temperature. To characterize the history of condition 110 up to a given instant, we can calculate the average gradient of this temperature (example: an average variation of 20°C / min "degrees Celsius per minute") during temperature transients (i.e., during transients inducing sudden temperature variations). Then, we can associate this average gradient with a count of such transients (for example: 2000 sudden temperature transients associated with an average gradient of 20°C / min from the time of installation up to said instant). Instead of the average of the historical values of the gradient of this temperature, one can also use any other statistical function such as median and standard deviation.
[0101] As another example, consider the case of a piece of equipment 2, a metal boiler tank, sensitive to and exposed to the temperature of the fluid it contains, and in particular subject to temperature cycling (i.e., subject to large temperature variations during heating or cooling, for example between 200°C and 80°C) and sensitive to these cycles. These cycles can then be counted (for example: 20 temperature cycles from installation to a given time) and this count of such cycles can be associated with the average of the cycling amplitudes in the history of said tank (for example: the average cycling amplitude of 150°C over the 20 cycles). Instead of the average of the historical amplitudes, any other statistical function, such as a median and a standard deviation, can also be used. When the invention chooses to characterize the history of a condition 110 by the history of a as a function of the physical quantity representing said condition 10, these functions may also include, but are not limited to, a count of said at least one fluctuation.
[0102] This is particularly the case when the equipment is sensitive to start-up or shutdown transients. Recording such transients then serves to characterize the usage history 11. Similarly, when the aging of equipment 2 continues during shutdown, recording the duration of the shutdown also serves to characterize the usage history. Likewise, to characterize the correlation 14 and in particular the state 13 of equipment 2 at a given time, in a way that is as specific as possible to the equipment 2 under consideration, the measured physical quantities or functions of these measured physical quantities characterizing the material state 13 of equipment 2 at said time are also determined.
[0103] For example, consider the case of a centrifugal pump 2. The state of the pump can then be characterized at any given time by means of the flow rate at that time and the discharge pressure at that time. At this stage, and by means of a technical analysis of the equipment 2, the invention has thus determined the critical manufacturing and maintenance tasks 90 specific to the equipment 2 in order to best characterize the manufacturing and maintenance history 9 of the equipment 2. Similarly, the operating conditions 110 specific to the equipment 2 have been determined, as well as the physical quantities or functions of physical quantities that best characterize said conditions 110 and the operating history 11 of the equipment 2. We have likewise determined the material indicators 120 that best characterize the state 13 of the equipment 2 at any given time, as well as the physical quantities or functions of physical quantities that best characterize said material indicators 120 at said time.
[0104] This technical analysis of equipment 2 thus makes it possible to determine the correlation 14 in a form specific to equipment 2. It follows that the determined correlation 14 is indeed specific to the physical phenomena occurring in each piece of equipment 2 in the series. Thus, according to one embodiment, in correlation 14, the functions of the measured physical quantities of the operating conditions 110 include a calculation of the time the measured physical quantities remain within at least one range of values; and / or a representative calculation of at least one fluctuation of the measured physical quantities; and / or a counting of said at least one fluctuation. Previously, in correlation 14, the invention chooses to characterize the manufacturing and maintenance history 9 of the equipment 2 (up to and including the last prior maintenance 10), by recording, for at least one (preferably each) critical manufacturing and maintenance task 90, the values that have been obtained These parameters are adopted during manufacturing and the various successive maintenance procedures (10) throughout the equipment's lifespan, up to and including the last maintenance procedure (10). The history of each critical task (90) is thus characterized by a list of values. For example, consider the aforementioned case of a pump-type piece of equipment (2) with a bearing, where the installation or replacement, as well as the type of pump bearing, corresponds to a critical task (90) in pump manufacturing and maintenance. In this example, the last maintenance procedure (10) (the one preceding the maintenance procedure to be performed) is considered to be the seventh (of rank k=7).The invention characterizes the history of this critical manufacturing and maintenance task 90 by the list [A,0,0,B,0,0,B,0], to translate the installation of the type A bearing at manufacturing 4, the replacement of the bearing with a new type B bearing at the third maintenance, the replacement of the bearing with a new type B bearing at the sixth maintenance, as well as the fact that no maintenance action was carried out on the bearing at the other maintenances.
[0105] The manufacturing and maintenance history 9 of the equipment 2 is thus a matrix grouping, for each critical task 90, the corresponding list of values. Furthermore, in order to reduce the matrix characterizing the manufacturing and maintenance history 9 (up to the last prior maintenance 10) to the necessary and sufficient information, and with a volume of information independent of the order of the last prior maintenance 10, the invention adopts the following principle, herein referred to as the "principle of residual parameters".According to one embodiment, in correlation 14, having reduced the manufacturing and maintenance history 9 to a critical task history 90 in the form of at least one list of successive values, in each list, only the residual value is chosen as the value adopted at the last maintenance at which the task 90 in question was performed, and only residual values are retained in the critical task history 90. In other words, in the manufacturing and maintenance history 9 and for each critical manufacturing and maintenance task 90: . - we consider the residual value of said task 90, as being the value adopted for said task 90, at the last manufacturing or maintenance stage at which said task 90 was carried out (at the previous manufacturing or maintenance stage 10); - we only keep the residual value of said task, in manufacturing and maintenance history 9.
[0106] In its final form, the manufacturing and maintenance history 9 is thus reduced to a list of values, that is, consisting of the residual values of the critical tasks 90 (i.e., one residual value per critical task 90). In so doing, the invention reduces the manufacturing and maintenance history 9 of the equipment 2, by characterizing said manufacturing and maintenance history 9 solely by the values of the critical tasks 90 that determine the behavior of the equipment 2 in operation following the last prior maintenance 10 considered. In practice, in the manufacturing and maintenance history 9: - the value characterizing each manufacturing and maintenance task 90 carried out during the last prior maintenance 10 overwrites the sequence of values characterizing the history of said task 90 (from manufacturing 4 up to the maintenance preceding the last prior maintenance 10); - of the manufacturing 4 and the prior maintenance 10 preceding the last prior maintenance 10, we keep, for each task 90 which has not been carried out at said last prior maintenance 10, only the value characterizing the last realization of said task 90. For example, we consider the case of a pump-type equipment 2 for which the task "installation or replacement of the pump bearing" and the task "installation or replacement of the pump impeller" are the only two critical manufacturing and maintenance tasks. If the history of the task "installation or replacement of the pump bearing" at the end of the seventh prior maintenance 10 is characterized by the list [A,0,0,B,0,0,B,0], the residual value for said task 90 is then "B", namely the value of task 90 adopted at the last maintenance where the task 90 in question was carried out (in this example, at the sixth maintenance). If the history of the task "installation or replacement of the pump impeller" at the end of the seventh prior maintenance 10 is characterized by the list [C,0,A,0,0,B,0,C], the residual value for said task 90 is then "C", namely the value of task 90 adopted at the last maintenance where the task 90 in question was carried out (in this example, at the seventh maintenance). The pump manufacturing and maintenance history 9 at the end of the seventh maintenance is then written [B;C], where the first term designates the residual value of the task "installation or replacement of the pump bearing" and the second term, the residual value of the task "installation or replacement of the pump impeller". In other words, the invention reduces the manufacturing and maintenance history 9 of equipment 2 (up to and including the last prior maintenance 10) to the configuration in which equipment 2 finds itself after the last prior maintenance 10. In doing so, the invention characterizes the manufacturing and maintenance history 9 (up to the last prior maintenance 10) by means of a vector whose size is independent of the order of said last prior maintenance 10. It should be noted that the invention can take into account the temporality of the values recorded, for example in the form of a data timestamp.
[0107] In the characterization of the manufacturing and maintenance history 9, this temporality can be translated by a time counter 91 attached to each residual value of critical task 90, said time counter 91 characterizing the duration since said task 90 has been carried out on equipment 2. At this stage, the correlation 14 previously determined in a form specific to equipment 2 (therefore specific to any type of equipment in the series), thanks to the technical analysis of equipment 2, is then determined in a format which lends itself to computer processing, in particular to learning by Machine Learning algorithms. Once the technical analysis of equipment 2 has been completed, the series-specific correlation 14 is identified. The following are then determined: - in the first term of correlation 14: the critical manufacturing and maintenance tasks 90 best characterizing the manufacturing and maintenance history 9 of equipment 2; - in the second term of the correlation 14: the physical quantities or functions of physical quantities best characterizing the conditions 110 and the history 11 of use of the equipment 2; - in the third term of the correlation 14: the physical quantities or functions of physical quantity best characterizing the material indicators 120 and the history 12 of the state 13 of the equipment 2.
[0108] Said critical tasks 90, physical quantities or functions of physical quantities identified in correlation 14, designate the raw data (measured and recorded for each piece of equipment 2 in the series) to be extracted initially from the histories 9, 11, 12 of the equipment 2. The invention therefore provides for compiling this accessible information from all the equipment 2 in the series under consideration. For all the equipment 2 in said series, the data associated with these tasks 90, as well as the data associated with the physical quantities or functions of physical quantities relating to the conditions 110 of use and the hardware indicators 120, as identified in said correlation 14, are retrieved and extracted so as to constitute triplets 140 of data and obtain a set 150 of raw data.For each piece of equipment 2 in the series and for each instant in their respective history 11, the three terms of the triplet 140 associated with the instant are: . - 1'set 95 of manufacturing and maintenance of said equipment 2, that is to say the fraction of the manufacturing and maintenance history 9 up to the last prior maintenance 10 preceding said moment; - 1'set 115 of use of said equipment 2, that is to say the fraction of the history 11 of use up to said instant; - the state 13 of said equipment 2 at said instant. As for the functions of physical quantities identified in the correlation 14, they indicate the conversions 141 to be performed subsequently on each of the triplets 140 of the set 150 of raw data, to constitute a final dataset 15 to be considered for modeling the correlation 14 via a virtual model 16. For example, we consider the case of a pump-type equipment 2 for which the terms of the correlation 14 specific to the series of said pump are thus characterized: - the manufacturing and maintenance history 9 is characterized by the critical tasks 90 relating to the pump bearing and impeller; - the usage history 11 up to the moment, is characterized by a first integral (INT1(t)) in time of the temperature over the duration between implantation and the moment and by a second integral (INT2(t)) in time of the suction pressure over the duration between implantation and the moment, as well as the average value of the temperature gradients (VMG(t)) over the duration between implantation and the moment, associated with the number of abrupt temperature transients (NTB(t)) over the same duration; - state 13 at the moment characterized by the values at the moment of flow rate and discharge pressure. The data to be extracted from the historical records 9, 11, 12 of each pump in the series under consideration, and for each moment in their respective historical record 11 of use, are then: - under the first term of correlation 14: the type of bearing and the type of impeller in place at the moment, to be extracted from the manufacturing and maintenance history 9; - under the second term of correlation 14: the values at the time considered of the physical quantities: temperature, suction pressure, to be extracted from the usage history 11; - under the third term of correlation 14: the values at the considered instant of the physical quantities: flow rate and discharge pressure, to be extracted from the state history 12. Once extracted, this data allows the construction, for each pump and for each instant (t) of their respective usage history 11, of the three terms of the triplet 140 associated with the instant, namely: - the set 95 of pump manufacturing and maintenance, that is to say, the fraction of the manufacturing and maintenance history 9, over the period between manufacturing 4 and the last prior maintenance 10 preceding said instant (t). The manufacturing and maintenance set 95 is, in the example, the history of tasks 90 (relating to the pump bearing and impeller) successively carried out during said period, or in other words, the matrix of two lists, one associated with the bearing and the other with the pump impeller, each list recording the successive values (type of bearing in place; type of impeller in place) characterizing the tasks 90 reviews made during that period; - the pump usage set 115, that is to say, in the example, the temperature and suction pressure values recorded from the time of installation until said instant (t); - the state 13 of the pump at said instant (t), that is to say, in the example, the flow rate and discharge pressure at said instant (t).
[0109] This yields as many triplets 140 as there are instants considered and as many pumps in the series: all these triplets 140 form the raw data set 150. In a second step, for each pump in the series and each triplet 140 associated with an instant, the pump manufacturing and maintenance data set 95 up to that instant is converted into a vector associated with that instant and indicating the residual value of each critical task 90. The corresponding vector is then, for each instant considered, [type of bearing in place at that instant; type of impeller in place at that instant], each term of the vector being associated with its own time counter 91. Similarly, for each pump in the series and each triplet 140 associated with an instant, the pump usage data set 115 up to that instant is converted.The physical quantity functions identified in correlation 14 being, in the example, the integrals (INT1(t),INT2(t)) in time of temperature and suction pressure, the average value of the temperature gradients (VMG(t)) associated with the number of abrupt temperature transients (NTB(t)), the conversions 141 to be carried out on the raw data, temperature and suction pressure, extracted from the usage history 11, are: . - the calculation of the integrals in time of temperature and suction pressure over the duration between implantation and the instant; - the calculation of the average value of temperature gradients over the same period and of the number of abrupt temperature transients over the same period. For each pump in the series and at each instant, the set 115 of pump usage between installation and instant is thus converted into a digital vector [INT1(t),INT2(t),VMG(t),NTB(t)] associated with instant (t). Finally, for each pump in the series and at each instant, the state 13 of the pump at instant is translated into a digital vector associated with instant [flow rate; discharge pressure]. Finally, each of the 140 triplets (associated with the different pumps (k) belonging to the series and the different times (t) in their respective usage history) is therefore converted into a vector [Mj(t)(k); Vu(t)(k); State(t)(k)], with: - Mj(t)(k), vector representing the set 95 of manufacturing and maintenance of the pump (k) up to the moment, in the form [bearing type; wheel type] each term of the vector being associated with its time counter 91; - Vu(t)(k), vector representing the set 115 of use of the pump (k) between the implantation and the instant, in the form [INTl(t),INT2(t),VMG(t),NTB(t)]; - State(t)(k), vector representing the state 13 of the pump (k) at the instant, in the form [flow rate; discharge pressure]. The raw data set 150 is thus converted into the final data set 15, namely: the set of vectors [Mj(t)(k); Vu(t)(k); Etat(t)(k)] for every pump (k) belonging to the series and for every instant (t) in their respective usage history 11. In short, system 1 plans to retrieve and process only the data considered relevant, identified in correlation 14. Optionally, system 1 can retrieve all the data relating to the histories 9, 11, 12 of the equipment 2, to extract and then process only the relevant data, as mentioned above. System 1 has thus converted the raw data set 150 into a data set 15, associated with the correlation 14 specific to the series under consideration. System 1 presents this dataset 15 as input to a machine learning project 142: at least one virtual model 16 is then trained on the basis of the dataset 15.Like correlation 14, the resulting virtual model 16 is specific to the series of equipment 2 considered. It also has the accuracy of empirical models (data-driven models).
[0110] Figures 5 and 6 illustrate the data collection and model learning process. In Figure 5, we consider the equipment 2 awaiting its maintenance 6 to be carried out, associated with its histories 9, 11, 12, as well as a given instant (t) during the period 5. Figure 5 represents the three terms of the triplet 140, linked by correlation 14, namely the manufacturing and maintenance set 95, the equipment use set 115, and the equipment state 13 at said instant (t), with: - the manufacturing and maintenance set 95 including manufacturing and maintenance decisions from rank 0 to rank k of the last maintenance preceding said instant (t); - the set 115 of usage including the sub-history 111 of usage during the period 5 up to audit instant (t); - the state 13 at time (t) of equipment 2. Figure 15 also represents the extraction and processing of data associated with equipment 2, showing in particular: - the extraction of data associated with set 95 of manufacturing and maintenance, set 115 of use and state 13 at time (t); - the constitution of the triplet 140 (set 95; set 115 of use; state 13 at said instant (t)) as visible in figure 5; - the formation of as many similar triplets as there are historical moments (t) at course of period 5; - the injection of said triplets into a set of 150 raw data, as seen in figure 5; - the constitution of said set 150 of raw data for each of the equipment of the same type as said equipment 2, within the installation 3.
[0111] Figure 6 schematically represents: - the recovery with transmission, with a view to their fusion, of the sets 150 of raw data extracted from several installations 3 operating the equipment of the same series as the equipment 2, followed by the conversions 141 carried out on these data, to constitute a set 15 of data; - the presentation of said data set 15 as input to a machine learning project 142 and the learning, on the basis of said data set 15, of a virtual model 16 modeling the correlation 14. By thus modeling the behavior of the equipment 2 (i.e. the behavior of the equipment in the series), the invention correlates, on the scale of the series, the observed behavior (namely the state 13) of each piece of equipment 2 to the manufacturing and maintenance history 9 and to the usage history 11 of said equipment 2. The invention is thus a form of exploitation of feedback, which derives its superior character from the functional relevance of the modeled correlation 14, in particular in that it is specific to the physical phenomena of which each piece of equipment 2 in the series is the site. According to one embodiment, the learning of model 16 falls within the domain of artificial intelligence and can be machine learning. In other words, the invention provides for the use of computer technology relating to artificial intelligence, in particular to the field of machine learning, which is based on mathematical and statistical approaches to give computers the ability to learn from data, that is to say, to improve their performance in solving tasks without being explicitly programmed for each one.
[0112] In particular, the supervisory system 1 includes a virtual model 16 resulting from such learning. Furthermore, the learning of model 16 can involve any type of machine learning, i.e., without limitation, supervised, semi-supervised, unsupervised, reinforcement learning, or transfer learning. Moreover, machine learning can implement learning methods of any category, said methods being able to be combined, i.e., without limitation: neural networks (including deep learning methods), k-nearest neighbors (KNN), genetic algorithms, genetic programming, or other methods such as, in particular, Bayesian networks, support vector machines (SVMs), Q-learning, decision trees, statistical methods, logistic regression, dis- analysis linear crime.
[0113] Preferably, the learning of the model 16 is a supervised, regression and deterministic machine learning problem (the model 16 preferentially determining a vector of quantitative and continuous data, from a set 15 of labeled data, a data vector which the invention has chosen to assimilate to the predicted state 130 of the equipment 2).
[0114] According to other embodiments, the model 16 can be of any type.
[0115] The invention aims to take into account the new manufacturing, maintenance, usage, and condition data generated since the first learning of model 16, within the framework of the machine learning project 142, and to update said virtual model 16. To do this, according to one embodiment, the preceding operations are repeated periodically: - We repeat the retrieval of new data from at least one manufacturer, maintenance provider and / or operator. Preferably, we retrieve the new data relating to all equipment in series 2 from all operators implementing the equipment in series; - we then perform the extraction and conversion 141 of said new data to obtain a complete dataset 15, followed by updating the learning of said model 16 on the basis of said completed dataset 15.
[0116] It should be noted that this update is carried out periodically, the periodicity being determined by the designer, manufacturer, maintainer and operators, according to criteria deemed most relevant by the latter or from the point of view of "data science". For example, a periodicity criterion could be a ratio of 10% of the time over which the raw dataset was compiled during the initial training. If the data was compiled over ten years, then the update period for the virtual model could be twelve months.
[0117] The invention thus allows for the exploitation of systematic feedback in a continuous or quasi-continuous manner.
[0118] According to the invention, once the correlation 14 has been determined and the learning of the model 16 has been carried out, the invention applies said model 16 to the equipment 2.
[0119] To this end, advantageously, during the maintenance 6 to be performed on said equipment 2, values are submitted to said model 16. These values correspond, on the one hand, to at least one of the tasks 90 of the manufacturing and maintenance history 9 and to at least one of the operating conditions 110 of the operating history 11. On the other hand, the values correspond to at least one of the tasks 90 of the maintenance 6 to be performed and to the said predicted operating conditions 110 of scenario 8. Indeed, model 16 has learned to correlate the state 13 of equipment 2 at a given time on the one hand the manufacturing and maintenance set 95 (i.e. up to the last prior maintenance 10 preceding said instant) and on the other hand the use set 115 up to said instant.
[0120] The model 16 is thus able to correlate the predicted state 130 of the equipment 2 at a given time, on the one hand with the modified manufacturing and maintenance history 9 of the maintenance tasks 90 to be carried out and on the other hand with the usage history 11 up to the audit time followed by the partial scenario 80.
[0121] In the application made of said model 16, the following input data are therefore submitted to said model 16: - an overall manufacturing and maintenance history consisting of the manufacturing and maintenance history 9 of equipment 2 (up to the last prior maintenance 10 preceding the maintenance 6 to be carried out), modified by the tasks 90 relating to the maintenance decision considered for the maintenance 6 to be carried out; - an overall usage history consisting of the usage history 11 of equipment 2 (up to the day before maintenance 6 to be carried out), followed by the partial scenario 80 considered (from the beginning of the planned operating period 70 until the moment considered during said planned period 70). Thus, the invention uses said model 16 as a simulator individualized to equipment 2. Indeed, model 16 allows for the simulation of the predicted state 130 of equipment 2 at a given time, firstly, for a given maintenance decision concerning the maintenance tasks 90 to be performed on said equipment 2, and secondly, for the subsequent use of said equipment 2 within the framework of scenario 8 up to that time. Furthermore, this simulation is specific to said equipment 2, because the manufacturing and maintenance history 9 up to the day before the maintenance 6 to be performed, and the usage history 11 up to the day before the maintenance 6 to be performed, which this simulation takes into account, are specific to the equipment 2 in question.
[0122] As can be seen in Figure 1, the application of the supervisory system 1 to the equipment 2 awaiting maintenance 6 to be carried out (of rank j), makes it possible to determine a predictive state 130 of said equipment 2 for a future time corresponding to the day before the predictive maintenance 7 (of rank j+1), taking into account the data relating to said equipment 2: - on the one hand, the manufacturing and maintenance history of said equipment and the usage history; - on the other hand, tasks 90 of a maintenance decision under maintenance 6 to be carried out and a usage scenario 8. Figure 1 specifically represents model 16 receiving, as input, the data aforementioned and generating, as output, the forecast state 130 of equipment 2 at the end of the forecast period 70. Further on, once these values are submitted, said model 16 generates a forecast state 130 of said equipment 2 subsequent to said maintenance 6 to be carried out. In other words, based on the input data, model 16 is able to predict a forecast state 130 of the equipment 2 for the current moment under consideration, that is, a value for the current moment under consideration for each of the material indicators 120 of this forecast state 130. This forecast state 130 is then compared to a minimum state 17 identified as required for the operation of said equipment 2. In other words, for each of the material indicators 120 of this forecast state 130, the value predicted by model 16 is compared to the minimum value required for the operation of the equipment 2 – for example, the minimum value of a performance of said equipment 2 (or, as the case may be, to the maximum value required for the operation of the equipment 2 – for example, the maximum value of wear of said equipment 2). If the predicted state 130 thus simulated for the moment considered satisfies the operating criterion (that is, if the value predicted by model 16 of each material indicator 120 of the predicted state 130 is greater (or less) than the minimum (or maximum) value required for the operation of equipment 2), then the predicted predicted state 130 corresponds to a state of good operation of equipment 2.
[0123] For example, we consider a case of equipment 2 of pump type, whose state 13 at an instant (t) can be characterized by means of the two material indicators 120, namely the flow rate (Q(t)) and the discharge pressure (Pref(t)), with respectively Qmin, and Prefmin for minimum values required for the operation of the pump. The minimum state 17 of the pump is therefore the set of values [Qm;n,Prefmin]. The operating criterion of the pump at time (t) is therefore that the predicted state 130 is greater than the minimum state 17, that is to say that the flow rate Q(t) and the discharge pressure Pref(t) satisfy the two conditions Q(t) > Qmin and Pref(t) > Prefmin. To identify whether the predicted state 130 of the pump, as predicted for the moment considered by the model 16, corresponds to a state of good operation of the pump, we then compare the state 130 to the minimum state 17, that is to say we compare the predicted values for the moment of the material indicators 120 flow rate Q(t) and discharge pressure Pref(t) to their respective minimum values Qmin and Prefmin.
[0124] Figure 7 illustrates, in the case of this example, the prediction of this forecast state 130 as well as the comparison to the minimal state 17. Figure 7 reproduces the previous example of a pump-type piece of equipment 2, whose state 13 at a given time (t) can be characterized by means of the two indicators 120 materials, namely the flow rate (Q(t)) and the discharge pressure (Pref(t)), with respectively as values of the indicators 120 minimum state materials 17, Qmin, and Pref min minimum values required for the operation of the pump. Figure 7 represents the evolution of the predicted state 130 of the pump (i.e. the successive values of the material indicators 120 Q and Pref), calculated by model 16, during the predicted period 70. Since the curves of the predicted state 130 over time are decreasing due to aging, Figure 7 also illustrates the comparison of the predicted state 130 at a time (t) with the minimum state 17: to identify whether the predicted state 130 of the pump corresponds to a state of good operation of the pump, we then compare the predicted state 130 to the minimum state 17, that is to say we compare the predicted values for said time of the material indicators 120 flow rate (Q(t)) and discharge pressure (Pref(t)) to their respective minimum values Qmin and Prefmin. In one embodiment, it is determined whether, for a given maintenance decision (decision concerning tasks 90 to be performed on equipment 2) under maintenance 6, followed by a given scenario 8, the maintenance decision will be sufficient, that is, whether it will give equipment 2 the potential for zero failures in operation, within the framework of said scenario 8 during and until the end of the planned period 70, i.e., until the day before the planned maintenance 7. To do this, in one embodiment, at least one variation of at least one of the values of tasks 90 of maintenance 6 is made. Then, when submitting the values to said model 16, the values of said variation are entered. Among all the variations, at least one sufficient maintenance decision is selected for the planned state 130 of said equipment 2, which is greater than or equivalent to the minimum state 17, at the time of said planned maintenance 7.
[0125] In other words, and in the case where the planned maintenance 7 corresponds to the next maintenance step following the maintenance 6 to be performed, a set of possible maintenance decisions (concerning the tasks 90) for the maintenance 6 to be performed is considered, and it is desired to determine, among these, the sufficient maintenance decisions. For example, a case of pump-type equipment 2 is considered, for which two maintenance tasks 90 are identified as critical: on the one hand, the replacement of the pump bearing (with types A and B as possible bearing types); on the other hand the replacement of the pump impeller (with types A, B and C as possible impeller types). The possible maintenance decisions for maintenance 6 to be carried out are then the twelve combinations of decisions formed from the following: - the three possible decisions regarding the landing: do nothing (leave the landing in place) or replace the landing with a type A or B landing; - the four possible decisions regarding the impeller: do nothing (leave the impeller in place) or replace the impeller with a type A or B or C impeller.
[0126] Among these twelve possible maintenance decisions, we wish to determine the sufficient maintenance decisions. Furthermore, if, among these possible decisions for maintenance 6 to be carried out, one maintenance decision is characterized by the vector [0;A] (i.e. bearing not replaced and impeller replaced by an impeller of type A) and another maintenance decision is characterized by the vector [0;0] (i.e. bearing not replaced and impeller not replaced), these two maintenance decisions for maintenance 6 to be carried out vary with regard to the impeller (in extenso, the variation of the values of tasks 90 of maintenance 6 to be carried out concerns the variation of the values "replacement" and "non-replacement" of the impeller).For each of the possible maintenance decisions for the maintenance 6 to be carried out, the aforementioned individualized simulation is implemented by model 16, to determine the predicted state 130 of equipment 2 at the end of the predicted period 70, i.e. on the eve of the predicted maintenance 7 (taking into account the use of equipment 2 which will follow the maintenance 6 to be carried out in the context of scenario 8 up to the time corresponding to the eve of the predicted maintenance 7, and taking into account the manufacturing and maintenance history 9 of equipment 2 up to the eve of the maintenance 6 to be carried out, and taking into account the usage history 11 of equipment 2 up to the eve of the maintenance 6 to be carried out).If, compared to the minimum state 17 required for the operation of equipment 2, the forecast state 130 (thus simulated by model 16 for the moment corresponding to the end of the forecast period 70), satisfies the criterion for the operation of equipment 2, the maintenance decision is considered sufficient (called "sufficient maintenance decision") to give equipment 2 the potential for zero failure in operation within the framework of scenario 8 until the end of the forecast period 70.
[0127] For example, consider the case of a pump-type piece of equipment 2 whose state 13 at a given time can be characterized by means of the two material indicators 120 of state 13, namely the flow rate (Q) and the discharge pressure (Pref), with (Qmin) and (Prefmin) respectively as the minimum values required for the pump to operate. The maintenance decision is considered sufficient if it allows a predicted state 130 (for the moment corresponding to the end of the predicted period 70) satisfying the pump operating criterion, namely the two conditions concerning the material indicators 120: (Q > Qmin) and (Pref > Prefmin) for the moment corresponding to the end of the forecast period 70. The simulation is repeated for each of the possible maintenance decisions (concerning tasks 90) for the maintenance 6 to be performed. The respective forecast states 130, thus simulated by model 16, currently corresponding to the end of the forecast period 70, allow us to sort the possible maintenance decisions for the maintenance 6 to be performed into sufficient and insufficient maintenance decisions. Thus, among all the variations of at least one of the values of tasks 90 for the maintenance 6 to be performed, at least one sufficient decision for said maintenance 6 has been selected, for the forecast state 130 of said equipment 2 that is greater than or equivalent to the minimum state 17, at the time of said forecast maintenance 7.According to one embodiment, for a given maintenance decision (concerning the tasks 90 of maintenance 6 to be carried out) and for each of the conditions 110 of use of scenario 8 of equipment 2 over the period 70 of the forecast, the maximum possible limit 18 which remains compatible with zero failure of equipment 2 (i.e. the limit compatible with uninterrupted operation of equipment 2, without risk of failure) is determined within the framework of scenario 8, during and until the end of the period 70 of the forecast, i.e. until the day before the next forecast maintenance 7 which follows the maintenance 6 to be carried out.
[0128] To do this, and for a given maintenance decision, a modification is made to the value of at least one of the predicted operating conditions 110 of scenario 8. When these values are submitted to said model 16, the values of said modification are entered along with the values of said maintenance decision. A limit is then calculated for at least one of said predicted conditions 110 for the predicted state 130 of said equipment 2 equivalent to the minimum state 17, at the time of the predicted maintenance 7.In other words, since scenario 8 of equipment 2 over the forecast period 70 is characterized by at least one forecast condition 110 of controllable use, a freedom of variation is allowed in the value of said forecast condition 110 of controllable use ("left free") and any other forecast conditions 110 of use (controllable and / or non-controllable) are set to their respective values in scenario 8. Model 16 is then used to solve, by iterative method and with respect to the forecast condition 110 of controllable use left free, the equation in which the forecast state 130 of equipment 2 at the end of the forecast period 70 corresponds to the minimum state 17. For a given maintenance decision (concerning maintenance tasks 90 of 6 to be performed) and for the considered predictable operating condition 110, a possible limit (preferably a maximum limit 18) has thus been determined which remains compatible with zero failure of equipment 2 (i.e., a limit compatible with uninterrupted operation of equipment 2, without risk of breakdown) until the end of the forecast period 70 (the other forecast usage conditions 110 remaining fixed at their respective values in scenario 8).
[0129] For example, in the case where the installation 3 is a truck and the supervised equipment 2 is the engine of said truck, we consider scenario 8 between the maintenance 6 to be carried out and the next planned maintenance 7, characterized by the following planned conditions 110 of use: - 110 predictable operating conditions: the average load to be transported (two tonnes = 2T), the distance to be covered (one hundred thousand kilometres = 100,000km), the average speed (ninety kilometres per hour = 90km / h); - 110 non-controllable forecast conditions of use: ambient air temperature (twenty degrees Celsius = 20°C), average slope of roads used (five percent = 5%). For this scenario 8, we want to determine the value of the maximum limit 18 of the average load to be transported, compatible with zero-failure until the planned maintenance 7 which follows the maintenance 6 to be carried out, the other planned conditions 110 of use remaining fixed at their respective target value of scenario 8. Using model 16 to solve, iteratively and with respect to the average load to be transported, the equation in which the predicted state 130 of equipment 2 at the end of the predicted period 70 corresponds to the minimum state 17, gives, as an example, a maximum limit 18 of 2.5T (for a target value of the average load to be transported of 2T in scenario 8). In Figure 8, we consider the aforementioned example of equipment 2 of the truck engine type for which the operating conditions 110 are the load to be transported P, the distance to be traveled d, the average speed V, the ambient air temperature T° and the average gradient of the roads used A. It is further assumed in Figure 8 that the predicted state 130 is defined by the material indicator 120 corresponding to the average compression pressure of the cylinders U, with the minimum value required for the proper functioning of equipment 2 being the value Umin of the minimum state 17.Figure 8 illustrates the calculation of the maximum limit 18 (Plim) of the controllable use condition 110 P, for equipment 2, associated with its fixed histories 9 and 11, for a given maintenance decision under maintenance 6 to be carried out, for a usage scenario 8 over the forecast period 70, defined by the respective target average values (Pc, de, Vc, T°c, Ac) of the controllable (P, d, V) and non-controllable (T°, A) use conditions 110 (Pc=2T; dc=100,000km; Vc=90km / h; T°c=20°C; Ac =5%, as a reminder in the aforementioned example). For the four other operating conditions (d, V, T°, A) calibrated to their respective average target values (dc = 100,000 km, Vc = 90 km / h, T°c = 20°C, Ac = 5%) of scenario 8, the maximum limit is determined (in the example, determined at Plim=2.5T) of condition 110 of controllable use (P), thanks to model 16 and iteratively. For this maximum limit 18 Plim, the forecast state 130 at the end of the forecast period 70 corresponds to the minimum required state 17, that is to say the state material indicator 120 then corresponds to the minimum required value (U=Umin).
[0130] Similarly, Figure 9 illustrates the calculation of the maximum limit 18 (dlim) of the controllable use condition 110 d. For the four other use conditions 110 (P,V,T°,A) set to their respective target average value (Pc=2T; Vc=90km / h; T°c=20°C; Ac =5%) of scenario 8, the maximum limit 18 (in the example determined at dlim=l 10,000km) of the controllable use condition 110 (d) is determined using model 16 and iteratively. For this maximum limit 18 (dlim), the forecast state 130 at the end of the forecast period 70 corresponds to the minimum required state 17, i.e., the material state indicator 120 then corresponds to the minimum required value (U=Umin). For each of the other forecast controllable usage conditions 110 of scenario 8, the same procedure is followed to determine the maximum possible limit 18 that remains compatible with zero equipment failure 2 (i.e.the limit compatible with uninterrupted operation of equipment 2 without risk of failure) within the framework of scenario 8, during and until the end of the planned period 70, namely until the day before the next planned maintenance 7 which follows the maintenance 6 to be carried out, and this for a given maintenance decision (concerning the tasks 90 of the maintenance 6 to be carried out).Thus determined, the said maximum limits 18 define the range of use compatible with zero failure of equipment 2 within the framework of scenario 8, during and until the end of the forecast period 70, and this for a given maintenance decision: it is known that zero failure of equipment 2 is possible as long as equipment 2 is operated, with a forecast condition 110 of controllable use kept less than or equal to the value of its maximum limit 18 thus determined, the other forecast conditions 110 of use remaining less than or equal to their respective target value of scenario 8.We can then graphically represent, on the one hand, the target values of the 110 forecast usage conditions of scenario 8 and, on the other hand, the values of the maximum limits 18 calculated previously for each of the 110 forecast usage conditions that can be controlled, using a spider diagram (“web mapping”), with each axis assigned to a forecast condition 110. We then obtain polygons with as many vertices as there are forecast conditions 110. We can thus superimpose a first polygon 180 called "target polygon 180" (representing the target value of each of the 110 forecast conditions of use of scenario 8), with a second polygon 181 called "limit polygon 181" (representing the maximum limit value 18 for each of the controllable predicted operating conditions 110 and the target value of scenario 8 for each of the non-controllable predicted operating conditions 110), the target polygon 180 is thus enclosed by the limit polygon 181. In Figure 10, we consider the aforementioned example of truck engine-type equipment 2, for which the operating conditions 110 are the load to be transported P, the distance to be traveled d, the average speed V, the ambient air temperature T°, and the average gradient of the roads used A. It is further assumed in Figure 10 that the predicted state 130 is defined by the material state indicator 120 corresponding to the average cylinder compression pressure U, with the minimum value required for the proper functioning of equipment 2 being the minimum state 17 value Uminde.Figure 10 shows the target polygon 180 and limit polygon 181, for equipment 2, associated with its fixed histories 9 and 11, for a given maintenance decision under maintenance 6 to be carried out, for a scenario 8 of use over the forecast period 70, defined by the respective target average values (Pc,dc,Vc,T°c,Ac) of the controllable (P,d,V) and non-controllable (T°,A) operating conditions 110 (Pc=2T; dc=100,000km; Vc=90km / h; T°c=20°C; Ac =5%, as a reminder in the aforementioned example). Figure 10 shows the "target polygon 180" in solid lines with, along the corresponding axes: - the respective target average values (Pc, de, Vc) of the 110 forecast controllable usage conditions (P, d, V) of scenario 8, represented by an open padlock; - the respective target average values (T°C, Ac) of the 110 forecast non-controllable operating conditions (T°, A) of scenario 8, represented by a closed padlock. Figure 10 shows the "boundary polygon" in dashed lines with, along the corresponding axes: - the values (Plim,dlim,Vlim) of the respective maximum limits 18 of the predictable controllable usage conditions 110 (P,d,V) of scenario 8, represented by an open padlock; - the respective target average values (T°c,Ac) of the predicted non-controllable operating conditions (T°,A) of scenario 8, represented by a closed padlock. According to one embodiment, the calculation of the maximum limits of the predicted operating conditions of scenario 8 is repeated for each of the maintenance decisions determined to be sufficient (i.e., the maintenance decisions concerning the tasks to be performed on equipment 2 during maintenance 6, which will give equipment 2 sufficient potential for zero failures in operation, within the framework of said scenario 8, during and until the end of the (planned period 70, i.e., until the day before planned maintenance 7). To do this, according to one embodiment, when submitting the values to said model 16, the selected values of said sufficient maintenance decision are entered. A maximum limit 18 of a planned usage condition 110 is calculated for this sufficient maintenance decision, for the planned state 130 of said equipment 2 equivalent to the minimum state 17, at the time of planned maintenance 7.
[0131] For each maintenance decision determined as sufficient for the maintenance 6 to be carried out and the scenario 8 considered, the maximum limits 18 of the predictable operating conditions 110 having been calculated, the target use (i.e. the target value of the predictable operating conditions 110 of scenario 8) is then represented graphically in a similar spider diagram, on the one hand, in the form of the target polygon 180 and, on the other hand, in the form of the limit use (i.e. the value of the maximum limits 18 of the predictable operating conditions 110 and the target value of the non-controllable predictable operating conditions 110 of scenario 8) is represented by the limit polygon 181, and this by superimposing the two polygons. According to one embodiment, a margin of use is determined for at least one of the 110 anticipated conditions of use of said scenario 8, as being the difference 182 between the corresponding value and the corresponding maximum limit 18.
[0132] Indeed, the invention quantifies, for all the 110 predictable controllable usage conditions of scenario 8, the gap 182 between their target value and the value of their respective maximum limit 18. In other words, the representation and superposition in the same spider diagram of the target polygon 180 of scenario 8 and the limit polygon 181 of the maximum limits 18, allows us to graphically represent, for each forecast condition 110 of controllable use, the difference 182 between the target value and the value of the maximum limit 18 of said forecast condition 110 of controllable use.
[0133] For example, in the aforementioned case of the truck engine, the difference 182 between the target value and the maximum limit value 18 is represented graphically as follows, for each predictable controllable operating condition 110: - for the average load to be transported: the difference 182 between the target value 2T and the maximum limit value 18 2.5T; - for the distance to be covered: the difference 182 between the target value 100,000km and the maximum limit value 18 110,000km; - for average speed: the difference 182 between the target value 90km / h and the maximum limit value 18 100km / h. Figure 10 represents, for each projected controllable use condition 110, the difference 182 between the target value and the maximum limit 18 value of said projected controllable use condition 110. Preferably, and in order to compare comparable differences 182, the axes of the cobweb diagram, on which the target polygon 180 of scenario 8 and the polygon 181 representing the limit of the maximum limits 18 are represented and superimposed, are normalized. In other words, for each projected controllable use condition 110, the target value and the value of its maximum limit 18 are represented on a scale, where each value is normalized (i.e., represented by the ratio between said value and the target value).
[0134] For example, in the aforementioned case of the truck engine and on a spider diagram whose axes are normalized, the deviations 182 between the target value and the value of the maximum limit 18 for each predicted controllable use condition 110 can thus be represented graphically: . - for the average load to be transported: the target value 2T is represented by the ratio 1 (2T / 2T), while the value of the maximum limit 18 is represented by the ratio 1.25 (2.5T / 2T), i.e. a difference of 25% between the two values; - for the distance to be covered: the target value is represented by the ratio 1 (100,000km / 100,000km), while the value of the maximum limit 18 is represented by the ratio 1.1 (110,000km / 100,000km), i.e. a difference of 10% between the two values; - For average speed: the target value is represented by the ratio 1 ((90km / h) / (90km / h)), while the maximum limit value is represented by the ratio 1.11 ((100km / h) / (90km / h)), i.e., a difference of 11% between the two values. Thus normalized, these differences between the target value and the maximum limit value can be compared: they allow us to identify the predictable operating conditions for which the difference between the target value and the maximum limit is significant or acceptable, as well as the predictable operating conditions for which the difference is not; which amounts to identifying the predictable operating conditions that tolerate a low or high, acceptable or unacceptable, margin before failure.
[0135] For example, in the aforementioned case of the truck engine, the difference 182 between the target value and the maximum limit value 18 is much greater for the average load to be carried (25%) than for the mileage to be covered and the average speed (10% and 11%, respectively). Consequently, the average load to be carried tolerates a greater difference 182 from its target value (without compromising zero breakdowns) than the mileage to be covered and the average speed. In other words, there is greater operational comfort before breakdown for the average load than for the mileage to be covered and the average speed, in order to benefit from zero breakdowns until the end of the forecast period 70. The operator should therefore be Much more vigilant, during the forecast period 70, in respecting the target values of scenario 8 for the mileage to be covered and the average speed, than for the average load transported. The invention therefore defines the margin of use before failure (MUAP) as the average of the normalized deviations 182 between the target value and the value of the maximum limit 18 for each forecast condition 110 of controllable use, or in other words, the average deviation between the target polygon 180 of scenario 8 and the polygon 181 limit of the maximum limits 18.
[0136] In the above example, the utilization margin before failure (MUAP), the average of the three normalized deviations 182 of 25%, 10% and 11%, amounts to approximately 15%. Thus, for a maintenance decision determined as sufficient for the maintenance 6 to be carried out and for a scenario 8, the maximum limits 18 of the controllable forecast conditions 110 having been calculated, the suitability of the maintenance decision (concerning the tasks 90 of the planned maintenance 6 to be carried out) to said scenario 8, as well as to the complete history of the equipment 2, is preferably quantified by means of the indicator of the utilization margin before failure (MUAP) associated with said maintenance decision.
[0137] For example, consider a case of equipment 2 of the truck engine type, for which seven sufficient maintenance decisions have been identified under maintenance 6 to be carried out, and for each of which the margin (MUAP) has been calculated, these decisions being: - a first decision (Dl) which replaces the diesel filter and allows a MUAP of 3% to be obtained; - a second decision (D2) which replaces the injectors and allows a MUAP of 7% to be obtained; - the third decision (D3) which replaces injectors and diesel filter and allows a MUAP of 9% to be obtained; - a fourth decision (D4) which replaces the injection pump and allows a MUAP of 11% to be obtained; - a fifth decision (D5) which replaces the injection pump and diesel filter and allows a MUAP of 13% to be obtained; - a sixth decision (D6) which replaces injection pump and injectors and allows a MUAP of 20% to be obtained; - a seventh decision (D7) which replaces the whole assembly (injection pump, injectors and diesel filter) and allows a MUAP of 25% to be obtained. For equipment 2 and its historical values 9 and 11, we observe that decision D7 proves more suitable for scenario 8 than the other decisions, particularly decision D1. Indeed, decision D1 allows for a MUAP of 3%: to benefit from zero downtime until the end of the projected period 70, an average value of deviations of 3% for a forecast usage condition of 110 should be met in relation to the target value of scenario 8. Decision D7 allows a MUAP of 25%: to benefit from zero failure until the end of the forecast period 70, a much larger average value of deviations of 25% for a forecast condition of use should be respected in relation to the target value of scenario 8.
[0138] The utilization margin (MUAP) thus makes it possible to quantify the suitability of each maintenance decision to the same predicted usage scenario 8. According to one embodiment, once the utilization margin is obtained, an optimal maintenance decision is selected from among the sufficient maintenance decisions, either as the sufficient decision enabling an acceptable utilization margin or as the sufficient decision enabling at least one of said acceptable deviations 182. In other words, for each of the decisions identified as sufficient for the maintenance 6 to be performed and for the scenario 8 considered, the maximum limits 18 of the controllable predictable operating conditions 110 having been calculated, and the suitability of the sufficient maintenance decision (concerning the tasks 90 for the planned maintenance 6 to be performed) to scenario 8 having been quantified by means of the utilization margin before failure (MUAP), the invention preferentially selects the optimal maintenance decision as the decision enabling the desired utilization margin.This preferentially determines the optimal decision regarding the volume of tasks 90 for maintenance 6 to be carried out. In short, this optimal decision is individualized to the complete history of equipment 2, and is adapted to scenario 8 (to allow zero failure until the end of the forecast period 70 as well as the desired margin of use before failure).
[0139] For example, we consider the aforementioned case of the truck engine, for which the seven sufficient maintenance decisions have been identified (from DI to D7) under maintenance 6 to be carried out, and for each of which the margin (MUAP) has been calculated. The operator may consider a difference of 10% between the target value of scenario 8 and the maximum limit value to be sufficient: - for the load to be transported (i.e. the operator estimates that the average load to be transported between maintenance 6 to be carried out and the next planned maintenance 7 will not exceed the target of 2T, beyond 10%); - for the mileage to be covered until the next maintenance (i.e. the operator estimates that the mileage to be covered between maintenance 6 to be carried out and the next planned maintenance 7 will not exceed the target of 100,000 kilometers, beyond 10%); - for average speed (i.e., the operator estimates that the average speed between the maintenance 6 to be carried out and the next planned maintenance 7 will not exceed the target of 90km / h, beyond 10%). The sufficient maintenance decision for maintenance item 6, with a margin (MUAP) above 10% and minimizing maintenance constraints, is therefore decision D4, which has a margin (MUAP) of 11%. In particular, decisions D5, D6, and D7 would also allow a margin (MUAP) greater than 10%, but, given the operator's preference, would constitute unnecessary over-maintenance with higher maintenance constraints. For the operator, the optimal maintenance decision for maintenance item 6 is therefore decision D4 (i.e., the one including the replacement of the injection pump). The invention also envisages selecting the optimal maintenance decision to be applied to the maintenance 6 to be carried out, as being the decision allowing the desired gap 182 between the target value of the scenario 8 and the value of the maximum limit 18, and this for one of the foreseeable conditions 110 of use.
[0140] For example, consider the aforementioned case of the truck engine, in which the seven sufficient maintenance decisions (DI to D7) have been identified under maintenance 6 to be carried out, and where the operator: - considers that the mileage covered on the eve of the next planned maintenance 7 cannot exceed the target of 100,000 kilometers of scenario 8; - considers that the average speed on the eve of the next planned maintenance 7 should not exceed the target of 90km / h of scenario 8; - wants a difference of at least 10% between the target value (2T) and the maximum limit value concerning the average load to be transported alone. In this example, and for each of the seven sufficient maintenance decisions, we then consider the difference 182 for the load to be transported between the target value in scenario 8 (2T) and the value of the maximum limit 18 calculated by the simulation of model 16: - the DI decision allows a 182% deviation of 5% for the load to be transported; - decision D2 allows a deviation of 182 of 11% for the load to be transported; - Decision D3 allows a 182% deviation of 15% for the load to be transported; - Decision D4 allows a 182% difference for the load to be transported; - Decision D5 allows a 21% deviation of 182 for the load to be transported; - Decision D6 allows a 182% difference of 33% for the load to be transported; - Decision D7 allows a 182% difference of 41% for the load to be transported. Given the assumptions and the operator's wishes, the optimal maintenance decision for maintenance 6 is, in this case, not decision D4, but decision D2 (with a difference of 11% between the target value and the maximum limit value for the load to be transported). In particular, the Decisions D3 to D7 would also increase the gap to over 10%, but, considering the operator's wishes, would constitute unnecessary over-maintenance, with greater maintenance constraints. Therefore, for the operator, the optimal maintenance decision for maintenance item 6 is D2 (i.e., the one including injector replacement).
[0141] The invention can further select the optimal maintenance decision to be applied to the maintenance 6 to be carried out, as being the decision allowing the desired average of the deviations between the target value of the scenario 8 and the value of the maximum limit 18, and this for a desired part of the anticipated conditions 110 of use.
[0142] For example, consider the aforementioned case of the truck engine, in which the seven sufficient maintenance decisions (DI to D7) have been identified under maintenance 6 to be carried out, - where the operator believes that the average speed on the eve of the next maintenance cannot exceed the target of 90 km / h, and - where the operator wants an average value of the deviations 182 of 10% (between target value of scenario 8 and maximum limit value 18) concerning the load to be transported and the mileage to be covered.
[0143] In this example, and for each of the sufficient maintenance decisions, we then consider the average of the deviations 182 between the target value of scenario 8 and the maximum limit 18 for the two predicted conditions 110 of use - the load to be transported and the mileage to be covered -, with the maximum limits 18 calculated by the simulation allowed by the model 16. In the example, it is further assumed that: - the DI decision allows for deviations of 5% and 2% respectively for the load to be transported and the mileage to be covered, i.e. an average deviation of 3%; - Decision D2 allows for deviations of 11% and 5%, i.e. an average deviation of 8%; - Decision D3 allows for deviations of 15% and 6%, i.e. an average deviation of 10%; - Decision D4 allows for deviations of 18% and 7%, i.e. an average deviation of 13%; - Decision D5 allows for deviations of 21% and 8%, i.e. an average deviation of 15%; - Decision D6 allows for deviations of 33% and 13%, i.e. an average deviation of 23%; - Decision D7 allows for deviations of 41% and 16%, i.e. an average deviation of 29%. Given the assumptions and the operator's preference, the maintenance decision for maintenance order 6 is, in this case, not decision D2 or decision D4, but decision D3, with an average deviation of 10%. Specifically, decisions D4 to D7 would also allow for an average deviation greater than 10%, but, considering the operator's preference, would constitute unnecessary over-maintenance, with increased maintenance constraints. For the operator, the optimal maintenance decision for maintenance order 6 is therefore, in this case, decision D3 (i.e., the one including the replacement of the injectors and the diesel filter). Thus, for equipment 2 which operated until maintenance 6 to be carried out and for scenario 8 to be executed after said maintenance 6 to be carried out, the simulation enabled by model 16 makes it possible to determine: - among a set of sufficient maintenance decisions, the optimal maintenance decision: this decision regarding the nature of the tasks 90 for maintenance 6 to be carried out is individualized to the complete history of the equipment 2, adapted to the scenario 8 to allow zero failure until the end of the forecast period 70 and allows the desired margin of use before failure (MUAP); - the associated limit use, namely the maximum possible use (compatible with zero failure until the due date of the next planned maintenance 7 following the maintenance 6 to be carried out), bounded by the maximum limits 18 of the planned controllable usage conditions 110 (the said maximum limits 18 being associated with the optimal maintenance decision).
[0144] For example, we consider the aforementioned case of the truck engine, for which the seven sufficient maintenance decisions (DI to D7) have been identified under maintenance 6 to be carried out, and in which the operator wants an average margin of use before failure (MUAP) of 10% for all the 110 predictable controllable operating conditions (average load to be carried, mileage to be covered until the next maintenance and average speed). In this example (and taking into account the manufacturing and maintenance history 9 and the truck engine usage history 11, as well as scenario 8 during the forecast period 70), among the sufficient maintenance decisions under maintenance 6 to be carried out, the optimal maintenance decision is decision D4 (i.e. the one including the replacement of the injection pump). The limit usage to be respected during the forecast period 70 to benefit from zero engine failure until the end of said forecast period 70 is the limit usage associated with this optimal maintenance decision (i.e. that bounded by the maximum limits 18 which were calculated for the forecast conditions 110 of use and for said maintenance decision D4: 2.3T with a deviation 182 of 18% from the target value of 2T, 107000km with a deviation 182 of 7% from the target value of 100,000km, 97 km / h with a deviation 182 of 8% from the target value of 90km / h).
[0145] In other words, given the method of calculating the maximum limits 18, zero downtime of the truck engine until the end of the forecast period 70 is possible as long as the engine is operated, with a given controllable condition 110 kept less than or equal to the value of its maximum limit 18 thus determined, the other conditions 110 being kept less than or equal to their respective target values of scenario 8. The construction of the limiting polygon 181, from the maximum limits 18 of the predicted conditions 110 of use, in order to quantify the average deviation with the target polygon 180 of scenario 8 (i.e. the margin of use before failure MUAP) is a conceptual tool provided for by the invention. Indeed, for the same target scenario 8 and in each case of sufficient maintenance decision 6 to be made, the limit polygon 181 and the margin (MUAP), always calculated according to the same methodology, allow us to quantify the suitability of each maintenance decision to the target scenario 8 (taking into account the manufacturing and maintenance history 9 and the equipment use history 11): the limit polygon 181 and the margin (MUAP) thus allow us to compare the maintenance decisions with each other.
[0146] Further, the invention provides for deducing all usage scenarios compatible with zero-failure until the due date of the next maintenance.
[0147] Indeed, given the method of calculating the maximum limits 18, the limit polygon 181 and its maximum limits 18 constrain the limit usage compatible with zero failure of the equipment 2 until the end of the forecast period 70, in the sense that zero failure is possible as long as the equipment 2 is operated, with a given forecast controllable usage condition 110 kept less than or equal to the value of its maximum limit 18, but only if the other forecast controllable usage conditions 110 are kept less than or equal to their respective target value of scenario 8. In particular, the limit polygon 181 does not allow us to deduce whether the use of the equipment 2, with two forecast controllable usage conditions 110 brought beyond their respective target value while respecting their respective maximum possible limit 18, remains compatible with the desired zero failure.For example, we consider the aforementioned case of the truck engine, with a target scenario 8, between the maintenance 6 to be carried out and the following planned maintenance 7, defined as follows: . - 110 predictable operating conditions: average load to be transported of 2T, distance to be covered of 100,000km, average speed of 90km / h; - 110 forecast non-controllable usage conditions: ambient air temperature of 20°C, average gradient of roads used of 5%.
[0148] It is further assumed in this example that the maximum limits 18 are for the average load to be transported of 2.5T, for the distance to be covered of 110,000km and for the average speed of 100km / h.
[0149] In this example, as constructed, the boundary polygon 181 indicates the following 8 usage scenarios as compatible with zero-failure until the next scheduled maintenance 7: - 2.5 T transported on average over 100,000 kilometres travelled on average at 90 km / h between the two maintenance visits; - 2 tons transported on average over 110,000 kilometers traveled on average at 90 km / h between the two maintenance periods; - 2 tons transported on average over 100,000 kilometers traveled on average at 100 km / h between the two maintenance operations. However, as constructed, the limiting polygon 181 does not allow us to determine whether transporting an average of 2.3 tons over 105,000 kilometers, at an average speed of 95 km / h between the two maintenance visits, is a usage scenario compatible with the desired zero breakdown rate until the end of the next scheduled maintenance visit 7. According to one embodiment, the invention therefore provides for determining, differently, all usage scenarios compatible with zero breakdown rate until the due date of the next maintenance visit. To this end, the system 1 includes a graphical representation in the form of a nomogram, with at least one curve indicating the maximum limit 18 of a first scheduled condition 110 as a function of at least a second of said scheduled conditions 110.This chart allows us to determine the usage scenarios compatible with zero-failure for the planned period 70 between the maintenance 6 to be carried out and the next planned maintenance 7. In other words, the invention aims to construct at least one nomogram-type graphical representation, for example in the form of a network of curves in an abacus, hereinafter "abacus", allowing for richer information regarding limit scenarios.For a given scenario 8 (defined by the target values of the planned controllable and non-controllable usage conditions 110), the constructed nomogram graphically represents the set of limit uses compatible with zero failure until the due date of the next planned maintenance 7, namely the set of n-tuples of the values of the planned controllable usage conditions 110, each n-tuple indicating, for the value of (n-1) given planned controllable usage conditions 110, the maximum limit 18 of the nth planned controllable usage condition 110 compatible with zero failure until the due date of the next planned maintenance 7 (the planned non-controllable usage conditions 110 of scenario 8 being locked to their respective values).For example, we consider the aforementioned case of the truck engine, with scenario 8 between maintenance 6 to be carried out and the following planned maintenance 7, the scenario defined by the following planned usage conditions 110: . - 110 predictable operating conditions: average load to be transported of 2T, distance to be covered 100,000km, average speed of 90km / h; - 110 forecast non-controllable usage conditions: ambient air temperature of 20°C, average gradient of roads used of 5%. In this case, and with a graphical representation of the mileage to be covered as a function of the average load to be transported, the nomogram is a network of curves, each associated with a given average speed between the two successive maintenance visits. Each iso-speed curve is the set of triplets of the values of the three predictable operating conditions (average load to be transported; distance to be covered; average speed), each triplet defining a limit usage compatible with zero breakdowns until the due date of the predictable maintenance visit (the non-controllable predictable operating conditions of scenario 8 - ambient air temperature and average gradient of the roads used - being locked at their respective values of 20°C and 5%).Thus, the curve associated with the average speed of 95 km / h allows us to deduce, for a given value of the average load to be transported (for example 2.3T), the value of the maximum limit 18 of the distance to be covered (101,000km in this example) until the next planned maintenance 7, i.e. the maximum possible value compatible with zero-breakdown until the deadline of the planned maintenance 7 (the planned non-controllable operating conditions 110 - ambient air temperature and average slope of the roads used - being locked to their respective values of the target scenario 8 (20°C and 5%)).To do this, in order to construct the curve associated with a given value of a predictable operating condition 110, the invention calculates, by iterative method and thanks to the simulation made possible by the model 16, the maximum limit 18 of a predictable operating condition 110, to be respected between the maintenance 6 to be carried out and the predictable maintenance 7, and this for different values of the remaining predictable operating conditions 110 (with the non-predictable operating conditions 110 always locked to their respective values of the target scenario 8). In the aforementioned example of the truck engine, there are three predictable operating conditions 110. The nomogram represents a network of curves, each associated with a value of a first predictable condition 110 (for example, average speed). To construct the curve associated with a value of said first predictable condition 110 (for example, an average speed of 100 km / h), the maximum limit 18 of a second predictable condition 110 (for example, the distance to be covered) is calculated using simulation by model 16, for different values of a third predictable condition 110 (in this example, the average load to be transported), with the non-controllable predictable operating conditions 110 (ambient air temperature and average gradient of the roads used) always locked to their respective values in the target scenario 8 (i.e., in this example, respect 20°C and 5%). Thus, for example, the nomogram designates the triplet (2.3T; 101,000km; 95km / h) as one of the possible limiting uses. In Figure 11, we consider the aforementioned example of equipment 2, a truck engine type, for which the operating conditions 110 are the load to be transported P, the mileage to be covered d, the average speed V, the ambient air temperature T°, and the average gradient of the roads used A. It is further assumed in Figure 11 that the predicted state 130 is defined by the material condition indicator 120 corresponding to the average cylinder compression pressure U, with the minimum value required for the proper functioning of equipment 2 being the value Uminde, the minimum state 17.Figure 11 shows an example of the use of such a nomogram representation, for equipment 2, associated with its fixed histories 9 and 11, for a given maintenance decision under maintenance 6 to be carried out, for a scenario 8 of use over the forecast period 70, defined by the respective target average values (Pc,dc,Vc,T°c,Ac) of the controllable (P,d,V) and non-controllable (T°, A) operating conditions 110 (Pc=2T; dc= 100,000km; Vc=90km / h; T°c=20°C; Ac =5% as a reminder in the aforementioned example).
[0150] In Figure 11, the visible nomogram is a network of three curves, in a graphical representation of the mileage to be covered (d) as a function of the average load to be transported (P). The visible curves are associated with three given average speeds between the two successive maintenance periods (90, 95, 100 km / h). For two desired target average values (Pc=2.3T; Vc=95km / h) for the controllable operating conditions P and V, the nomogram, as calculated by the model 16, indicates the value of the maximum limit 18 (dlhn= 101,000km) for the last controllable operating condition (d).Thus the usage scenario, defined by the three desired target average values (Pc=2.3T, dlhn= 101,000km and Vc=95km / h) for the 110 forecast controllable usage conditions and the two target values (T°c=20°C and Ac =5%) for the two 110 forecast non-controllable usage conditions, corresponds to a limit scenario compatible with the desired zero-failure until the end of the forecast period 70, i.e. compatible with a forecast state 130 of the equipment 2 at the end of the forecast period 70 corresponding to the minimum required state 17 of good operation (U = Umin). . According to one embodiment, once maintenance 6 has been carried out and during the forecast period 70, the invention updates, at any time, the forecast state 130 of the equipment 2 as it will be at the end of the forecast period 70, i.e., on the eve of the next forecast maintenance 7. In other words, at any given time during the forecast period 70, when scenario 8 of the equipment 2 is being executed, it is determined whether, taking into account the actual use of the equipment 2 since the beginning of the forecast period 70, said equipment 2 The device maintains, at the given moment, sufficient potential for zero downtime during operation, within the framework of the remaining timeframe of said scenario 8, throughout the remaining timeframe and until the end of the forecast period 70, i.e., until the day before the next forecast maintenance 7 following the maintenance 6 to be performed. The invention already provides for selecting, on the eve of the maintenance 6 to be performed, the optimal maintenance decision from among sufficient maintenance decisions, which allow for the forecast state 130 at the end of the forecast period 70 (and taking into account a scenario 8) compatible with the operation of the equipment 2. However, the invention provides for updating, at any time during the forecast period 70, the forecast state 130 as it will be at the end of the forecast period 70, because this updating is considered potentially necessary. Indeed, the actual use of equipment 2, during the elapsed part of the planned 70-day operating period between its start and the time considered, may have differed from the planned use of scenario 8.
[0151] Furthermore, during the elapsed portion of the predicted period 70, the equipment 2 may have experienced abnormal operating transients, including incursions into the near-destruction domain, or incursions into the intermediate domain between said near-destruction domain and the normal operating domain. The invention assumes that the aging, and therefore the future behavior, of the equipment is impacted by the accumulation of these abnormal transients. The invention therefore provides for updating, during operation, the predicted state 130 as it will be at the end of the predicted period 70, with the accumulation of any abnormal transients that have occurred since the beginning of the predicted operating period 70 up to the present moment.Similarly, if the invention provides, on the eve of maintenance 6 to be carried out, for determining the optimal decision for the maintenance 6 to be carried out as well as the associated limit use (namely the maximum possible use to be respected in operation, at the end of said maintenance 6 to be carried out, for zero failure until the due date of the next planned maintenance 7), the invention provides for the updating in operation of this limit use, as being potentially necessary. Indeed, the limit of equipment 2 is likely to contract, in operation, due to actual use of equipment 2 (and more constraining for equipment 2 than the scenario 8 initially planned), during the past part of the forecast period 70, and due to possible abnormal operating transients.
[0152] By updating the operating usage limit compatible with zero-failure until the next scheduled maintenance 7, the invention provides for enabling the user to thus identify the maximum limits 18 up to which he can push the use of the equipment without compromising zero breakdowns or in which it must restrict the use initially envisaged to benefit from zero breakdowns. For example, consider the aforementioned case of the truck engine, where the maintenance decision D6 to be carried out (namely the replacement of the diesel filter and the injection pump) was adopted, and where, at the end of half of the planned period 70, it turns out that the average load carried does indeed meet the target value of 2T, that the mileage covered (50,000km) does indeed meet the pro rata of the target value of 100,000km at the end of the planned period 70, but that the average speed was 105km / h, exceeding both the target value of 90km / h of scenario 8 and the value of the maximum limit 18 of 100km / h. Under these conditions (and in addition to any abnormal transients during the elapsed portion of the forecast period 70), continuing to operate the truck under the initial scenario 8 with its target values (2T, 100,000km, 90km / h) will not be compatible with zero breakdowns until the next scheduled maintenance 7. Therefore, the calculation of the maximum limits 18 of the various forecast operating conditions 110 must be updated. The operator must adhere to these limits to achieve zero breakdowns until the end of the forecast period 70. To do this, after the aforementioned maintenance 6 has been completed, updated values are submitted to the model 16. These values correspond, firstly, to at least one of the tasks 90 from the manufacturing and maintenance history 9, as before, and secondly, to at least one task 90 from the maintenance 6 that has been performed. These values also correspond to at least one of the usage conditions 110 of the usage history 11 since the aforementioned maintenance 6 was performed. These values also correspond to the predicted usage conditions 110 of scenario 8. Therefore, based on these values, said model 16 updates the predicted state 130 of said equipment 2. Furthermore, said predicted state 130 is compared to the minimum state 17 identified as required for the operation of said equipment 2. In short, according to one embodiment, after said maintenance 6 has been performed, said model 16 is subjected to values - on the one hand, of at least one of the tasks 90 of the manufacturing and maintenance history 9, of at least one task 90 of the maintenance 6 performed, and of at least one of the usage conditions 110 of the usage history 11 since said maintenance 6 was performed - on the other hand, the said conditions 11 forecasting the use of scenario 8, the said model 16 updating the forecast state 130 of said equipment 2, said forecast state 130 being compared to a minimum state 17 identified as required for the operation of said equipment 2. In other words, at a given moment during the next forecast period 70, the prediction of the forecast state 130 is updated visualization of equipment 2 for what it will be at the end of the planned period 70, i.e., on the eve of the planned maintenance 7. To do this, and because at the time considered, maintenance 6 has been carried out and equipment 2 is in operation, we use said model 16 as a simulator individualized to equipment 2, adapting the arguments submitted as input to model 16, namely: - the manufacturing and maintenance history of equipment 2; - tasks 90 performed under maintenance 6; - the usage history 11 up to the time considered, followed by the usage during the elapsed part of the planned operating period 70 up to the time considered; - the use of equipment 2 planned in the remainder of scenario 8, from that moment and for the remainder of the planned operating period 70 until the day before the next planned maintenance 7. This possibility of being able to update the planned state 130 of equipment 2, as it will be at the end of the planned operating period 70 after the upcoming execution of the entirety of scenario 8 (taking into account the tasks 90 actually performed for maintenance 6 and taking into account the actual use made of installation 3 and equipment 2 from the beginning of the planned operating period 70 until the moment considered) therefore allows the various applications of model 16 as the invention provides for them on the day before the maintenance 6 to be carried out. In particular, the use of model 16 makes it possible to determine, at any time during the forecast period 70, whether the continuation of scenario 8 remains compatible with zero failure until the end of the forecast period 70, by comparing said forecast state 130 to the minimum state 17 required for the operation of equipment 2. In particular, the use of model 16 allows for the updating, at any time during the forecast period 70, of the maximum usage limits 18 compatible with zero downtime until the next scheduled maintenance 7, the limit polygon 181, the maximum operating margin before failure (MUAP), and the nomogram of all limit scenarios. This allows the user to identify the limits to which they can push the use of equipment 2 without compromising zero downtime, or within which they must restrict the initially planned use to achieve zero downtime. The invention updates, at a given time during the forecast period 70, the forecast state 130 of equipment 2 as it will be at the end of said forecast period 70 and after the execution of scenario 8.Similarly, the invention updates, at a given moment during the forecast period 70, the forecast state 130 of the equipment 2 as it will be at the end of the forecast period 70 and after the execution of a scenario different from the initially planned use scenario 8, with regard to the period between said moment and the end of the . forecast period 70.
[0153] Among the arguments to be submitted as input to model 16, instead of submitting the use of equipment 2 planned within the framework of scenario 8, from said moment and for the remainder of the planned operating period 70 until the day before planned maintenance 7, we submit to model 16 the use of equipment 2 within the framework of the unforeseen scenario, from said moment and for the remainder of the planned operating period 70 until the day before planned maintenance 7.Similarly, the invention updates, at a given moment during the forecast period 70, and with a view to the execution of a scenario different from the initially planned use scenario 8, between said moment and the end of the forecast period 70, the maximum limits 18 of the limit use compatible with zero-failure until the next planned maintenance 7, the limit polygon 181, the margin of use before failure (MUAP) as well as the nomogram of all the limit scenarios: which allows the user to identify the limits to which he can push the use of the equipment without compromising zero-failure or within which he must restrict the initially planned use to benefit from zero-failure and this with a view to the execution of a scenario different from the initially planned use scenario 8.Thus the invention allows the operator to decide, during the forecast period 70, how he can or should use the equipment 2, to keep this use compatible with the zero-failure requirement until the next maintenance deadline, in the context of an initially planned scenario 8 as well as in the context of an unforeseen scenario, replacing the initially planned scenario 8.
[0154] Advantageously, the simulation using model 16 also makes it possible to determine, step by step, the sufficient maintenance decisions for the successive planned maintenance 7s during the life of the equipment 2, taking into account the target use (namely the successive scenarios 8) desired by the operator in both the short and long term. System 1 thus makes it possible to determine at any time the various possible sufficient maintenance programs throughout the life of the equipment 2 for a desired usage profile. To do this, according to one embodiment, at least the following steps are performed. Firstly, we assume that at least the said sufficient decision of maintenance 6 to be carried out has been made and we assume that scenario 8 is executed until the planned maintenance 7 which follows said maintenance 6 to be carried out. Next, at least one variation of at least one of the values of tasks 90 of said planned maintenance 7 is made. When submitting the values to said model 16, the values of said variation are entered along with the values of the following scenario 81 planned for the next planned period 710 (i.e., the operating period). included between said planned maintenance 7 and the following planned maintenance 71). Among all the variations, at least one sufficient decision of said planned maintenance 7 is selected, for the planned state 130 of said equipment 2 greater than or equivalent to the minimum state 17, at the time of the planned maintenance 71 following said planned maintenance 7.
[0155] Once this sufficient decision has been selected, these steps are repeated for any subsequent planned maintenance in the equipment's life. In short, a series of sufficient maintenance decisions is determined on a recurring basis for planned maintenance 70 subsequent to the maintenance 6 to be performed, once the sufficient maintenance decisions for the maintenance 6 to be performed have been defined. Indeed, on the eve of the maintenance 6 to be performed (given a maintenance decision determined as sufficient for the maintenance 6 to be performed and assumed to have been carried out, given the scenario 8 assumed to have been carried out), the model 16 makes it possible to determine the sufficient maintenance decisions for the next planned maintenance 7 that will follow the maintenance 6 to be performed.In this sense, model 16 allows for recurrence, to determine step by step, from one planned maintenance 7 to the next, the sufficient maintenance decisions.
[0156] To this end, and as before, for each of the possible maintenance decisions for the planned maintenance 7 following the maintenance 6 to be performed, the planned state 130 of equipment 2 is simulated for what it will be at the end of the next planned period 710, following a subsequent scenario 81, i.e., on the eve of the next planned maintenance 71 (i.e., following the planned maintenance 7 under consideration). To do this, and taking into account these assumptions, the aforementioned model 16 is used as a simulator individualized to equipment 2 by adapting the arguments submitted as input to model 16, namely: - the manufacturing and maintenance history 9 of equipment 2 (up to the day before maintenance 6 to be carried out), modified from the tasks 90 assumed to have been carried out as part of maintenance 6 to be carried out; - tasks 90 of the possible maintenance decision considered under the provisional maintenance 7; - the usage history 11 (up to the day before maintenance 6 to be carried out), followed by scenario 8 assumed to be executed during the forecast period 70; - the following scenario 81 of equipment 2 (during the next forecast period 710 and up to the day before the next forecast maintenance 71). If, compared to the minimum state 17 required for the operation of equipment 2, the forecast state 130 of equipment 2, thus simulated by model 16 for the corresponding time At the end of the following planned period 710 (i.e., on the eve of the next planned maintenance 71), the equipment 2 operating criterion is met. The maintenance decision for the planned maintenance 7 following the scheduled maintenance 6 can then be considered sufficient to give equipment 2 the potential for zero failures during operation under the following scenario 81, throughout and until the end of the next planned period 710. Recurrence is thus established.Indeed, on the one hand, on the eve of maintenance 6 to be performed (maintenance of rank j) (given a maintenance decision determined as sufficient for maintenance 6 to be performed and assumed to have been carried out, and given a scenario 8 assumed to be carried out during the following planned operating period 70), the model 16 made it possible to determine the maintenance decisions sufficient for the planned maintenance 7 (maintenance of rank j+1), namely the decisions allowing for zero failures for a subsequent scenario 81, during and until the end of the following planned period 710. On the other hand, the invention makes it possible to determine the decisions sufficient for maintenance 6 to be performed.The simulation using model 16 therefore makes it possible to determine, by recurrence, step by step, the sequences of sufficient maintenance decisions for any planned maintenance 7 up to the desired rank of planned maintenance 7 in the life of equipment 2 (maintenance of rank j+n). In figures 12 and 13, we consider equipment 2 associated with its histories 9 and 11, with (in figure 12) a scenario 8 of use during the planned period 70, followed (in figure 13) by a subsequent scenario 81 during the following planned period 710.Figures 12 and 13 illustrate the step-by-step simulation of the 130 forecast states of equipment 2 on the eve of each 7,71 forecast maintenance and the determination by recurrence of sufficient maintenance decisions for said 7,71 forecast maintenance: model 16 allows us to determine sufficient maintenance decisions for a 6 maintenance to be carried out (of rank j), model 16 then allows us to determine sufficient maintenance decisions for the next 7 forecast maintenance of rank j+l.Initially, Figure 12 illustrates that, for a possible maintenance decision under maintenance 6, model 16 allows us to predict the forecast state 130 of equipment 2 at the end of the forecast period 70: in particular, model 16 therefore allows us to determine whether this possible decision is sufficient to achieve the desired zero failures until the end of the forecast period 70 (depending on whether the forecast state 130 of equipment 2 at the end of the forecast period 70 is sufficient with regard to the minimum state 17). Model 16 effectively determines the sufficient maintenance decisions, among the possible maintenance decisions, under the [specific maintenance requirement]. maintenance 6 to be carried out. In a second step, figure 13 illustrates that, for a maintenance decision identified as sufficient for a maintenance of rank j (in this case for the maintenance 6 assumed to have been carried out) and for a possible maintenance decision for the following maintenance of rank j+1 (in this case for the following planned maintenance 7), model 16 makes it possible to predict the planned state 130 of equipment 2 at the end of the following planned period 710: model 16 therefore makes it possible to determine if this last possible decision is sufficient to allow the desired zero failure until the end of the following planned period 710 (depending on whether the planned state 130 of equipment 2 at the end of the following planned period 710 is sufficient with regard to the minimum state 17).Figure 13 illustrates that, for a maintenance decision identified as sufficient for maintenance of rank j, model 16 makes it possible to determine the sufficient maintenance decisions, among the possible maintenance decisions, for the next maintenance of rank j+1.
[0157] As can be seen in Figures 12 and 13, the system 1 thus allows the recursive determination of sufficient maintenance decisions from one maintenance to the next in the life of the equipment 2, that is, step by step, from a maintenance of rank k to a maintenance of rank k+1. In particular, Figures 12 and 13 show how the model 16 is used, on the one hand, to determine the predicted states 130 of the equipment 2 during the predicted period 70 and, on the other hand, to determine the predicted states 130 of the equipment 2 during the following predicted period 710. Figures 12 and 13 illustrate the construction of the input arguments of the model 16 in both cases.In particular, the step-by-step simulation of the predicted states 130 on the eve of each predicted maintenance 7 is based on the fact that the simulations of the predicted states 130 for the following predicted period 710 after the maintenance of rank j+1 (as seen in Figure 13) incorporate the maintenance decision that the simulations have identified as sufficient for the maintenance of rank j (as seen in Figure 12). Similarly, the invention determines, step by step, the successive predicted states 130 of the equipment 2 during the two predicted periods 70,710. For this, we consider, on the one hand, two maintenance decisions respectively for maintenance 6 to be carried out (maintenance of rank j) and for planned maintenance 7 (maintenance of rank j+1) and, on the other hand, two usage scenarios 8,81 for each of the said planned periods 70,710.In this respect, figure 14 represents an example of two curves of the predicted state 130 of equipment 2 of pump type (reduced in this example to its material indicator 120 “flow Q”).
[0158] Model 16 makes it possible to predict the predicted state 130 of the pump, for the different times during the planned period 70 (from maintenance order j to maintenance order j+1) and then for the different times during the following planned period 710 (from maintenance order j+1 to maintenance order j+2). In particular, the simulation of model 16 highlights that it will be necessary to intervene on the pump at maintenance order j+1 to achieve the desired zero-failure rate (otherwise, the pump would fail before maintenance order j+2, as shown by the dashed extension of the curve).
[0159] Figure 14 notably highlights a drop in flow rate on either side of maintenance order j+1, due to a maintenance decision made on the pump (at the due date of said maintenance order j+1). This maintenance decision at order j+1 is also sufficient because it ensures sufficient flow rate until the day before maintenance order j+2. Similarly, and more generally, the invention determines, step by step, the successive predicted states 130 of the equipment 2 as they will be during as many successive predicted periods 70,710 as one wishes to consider in the life of the equipment 2.To this end, we consider, on the one hand, a series of maintenance decisions respectively for maintenance 6 to be carried out (maintenance of rank j) and then for the various planned maintenance 7,71 preceding the planned periods 70,710 considered (maintenance of rank j+1 to j+k) and, on the other hand, a series of usage scenarios 8,81 for each of the planned periods 70,710 considered. In this respect, Figure 15 represents an example of two overall curves of the planned state 130 of the pump-type equipment 2 (reduced in this example to its material indicator 120 "flow rate Q").
[0160] Model 16 makes it possible to predict the predicted state 130 of the pump for the different times of the period between maintenance of rank j and maintenance of rank j+k and for the different times of the period between maintenance of rank j+k and maintenance of rank j+n. These curves result from the prediction, through the simulation of model 16, of the successive predicted states 130 of the pump for each operating period between two successive maintenances (intervals delimited by the vertical dashed lines in Figure 15).
[0161] It is thus demonstrated that it will be necessary to intervene on the pump at maintenance interval j+k to achieve the desired zero-failure rate during each operating period between maintenance interval j and maintenance interval j+n. Similarly, Figure 15 notably highlights a drop in flow rate on either side of maintenance interval j+k, due to a maintenance decision made on the pump (at the end of said maintenance interval j+k). The invention thus makes it possible to determine the predicted state 130 of the equipment 2 as it will be at any time during any subsequent predicted period 710 in the equipment life 2 (within the framework of a given scenario 81 and taking into account a maintenance decision determined to be sufficient for the planned maintenance 7).
[0162] Thanks to this predictive capability, and as before, the invention makes it possible to identify, for each subsequent forecast period 710 in the life of the equipment 2, the limits to which the use of the equipment can be pushed without compromising the desired zero failure or within which the initially envisaged use must be restricted to benefit from zero failure.
[0163] Indeed, for each subsequent forecast period 710 in the life of the equipment 2 (at the end of a given subsequent scenario 81 and taking into account a maintenance decision determined to be sufficient for the corresponding forecast maintenance 7), the invention: - determines, as before, the maximum limits 18 of the foreseeable controllable usage conditions 110; - thus limits the usage compatible with zero-failure during the following scenario 81 during the said following planned period 710 and this until the day before the next planned maintenance 71; - these maximum limits 18 of the forecast conditions 110 of use being thus calculated, quantifies, as before and preferentially, the suitability of said sufficient maintenance decision (concerning the tasks 90 for said forecast maintenance 7) to the following scenario 81, by means of the utilization margin before failure (MUAP) or by means of the gap 182 between target value of the following scenario 81 and maximum limit 18 for at least one forecast condition 110 of controllable use of said following scenario 81; - constructed, as before, the nomogram indicating all the valid limit scenarios for the following forecast period 710. Thus, the simulation by model 16 makes it possible to determine, recursively, step by step, the sequences of sufficient maintenance decisions for any forecast maintenance 7, up to the desired rank of forecast maintenance 7. Furthermore, for each of the following forecast periods 710, the simulation enabled by model 16 makes it possible to determine, for each of said sufficient maintenance decisions, the maximum limits 18 for each forecast condition 110 of controllable use of the corresponding scenario 81, the limit polygon 181, the margin of use before failure (MUAP), and the nomogram of limit scenarios. According to one embodiment, the optimal decision is selected from among at least the sufficient decision for the corresponding maintenance. In other words, for each planned maintenance step 7 in the life of the equipment 2, the decision is selected, as before for the maintenance 6 to be carried out. optimal maintenance, from among the maintenance decisions determined to be sufficient. Said optimal decision is selected, preferably with regard to the utilization margin before failure (MUAP) permitted by each sufficient maintenance decision considered, or with regard to at least a difference 182 between the target value of scenario 8 and the value of the maximum limit 18 for at least one of the 110 forecast conditions of use of said scenario 8.
[0164] Thus, the invention provides for optimizing each of the planned maintenance 7 during the life of the equipment 2, by determining the optimal maintenance decision for each maintenance stage during the equipment 2's life and considering successive scenarios 8 during the equipment's life. This decision regarding the nature of the tasks 90 for the planned maintenance 7 is individualized to the equipment's complete history, is adapted to scenario 8 to enable zero downtime until the end of the planned period 70, and allows for the maximum allowable uptime (MUAP) desired by the operator. In other words, the invention recursively determines the sequence of optimal maintenance decisions for each of the planned maintenance 7 during the remaining part of the equipment 2's life.Furthermore, for each of the 70 projected periods in the equipment's life, the invention determines the maximum limits (for each projected controllable use condition of the corresponding scenario), the limit polygon, the maximum allowable use (MUAP), and the limit scenario chart, taking into account the maintenance decisions identified as optimal. For each of the 70 projected periods in the equipment's life, the invention thus makes it possible to identify the limits to which the equipment can be used without compromising zero failure, or to which the initially planned use must be restricted to achieve zero failure. In other words, the invention recursively determines the sequence of limit uses of the equipment for each of the 70 projected periods in the remainder of the equipment's life.
[0165] As seen previously, for the maintenance 6 to be performed, model 16 makes it possible to distinguish between a sufficient maintenance decision and an insufficient maintenance decision. In the case of a sufficient maintenance decision, the maintenance decision is suitable for the usage scenario 8 in that it is sufficient to allow for zero downtime until the end of the forecast period 70. In the case of a maintenance decision determined to be insufficient, the operation of equipment 2 within the framework of the usage scenario 8 leads to the failure of equipment 2 before the end of the forecast period 70. In this latter case, the calculation of the failure deadline 19 is then valid. Thus, according to one embodiment, when a maintenance decision is insufficient with a forecast state 130 (at (during the forecast period 70) lower than the minimum state 17, the failure deadline 19 for the said insufficient maintenance decision is then determined. The failure deadline 19 corresponds to the moment when the forecast state 130 is equivalent to the minimum state 17. Simulation using model 16 allows this failure deadline 19 to be determined. To do this, and for a given insufficient maintenance decision concerning the maintenance 6 to be carried out, for the forecast period 70 and its scenario 8, the following procedure is followed. A given moment during the forecast period 70 is considered, along with the partial scenario 80 of scenario 8 corresponding to said moment. Model 16 is used as a simulator individualized to equipment 2, adapting the arguments submitted as input to model 16, namely: - the manufacturing and maintenance history 9 of equipment 2; - the insufficient maintenance decision considered under maintenance 6 to be carried out; - the history 11 of equipment usage 2; - the use of equipment 2 planned within the framework of the partial scenario 80 defined up to the considered time in the forecast period 70. Model 16 is then used to solve, with respect to said time, the equation in which the forecast state 130 of equipment 2 at said time corresponds to the minimum state 17. For the insufficient maintenance decision considered and for scenario 8, the failure time 19 is determined to be the solution of said equation.
[0166] For an insufficient maintenance decision, considered under maintenance 6 to be performed, the equipment 2 failure time 19 is thus determined. The same procedure can be used to determine, during the following forecast period 710, the failure time 19 associated with an insufficient maintenance decision under forecast maintenance 7 and associated with the following scenario 81. Figure 16 considers pump-type equipment 2, associated with its histories 9 and 11, a scenario 8 during the forecast period 70 (between maintenance 6 to be performed (of rank j) and the following forecast maintenance 7 (of rank j+1)), a partial scenario 80 of scenario 8 and associated with a time point, as well as an insufficient maintenance decision considered under maintenance 6 to be performed.Figure 16 represents the curve of the predicted state 130 of the pump (reduced in this example to its material indicator 120 "flow Q") which model 16 allows to be predicted for the different times of the predicted period 70. The curve of the predicted state 130 over time is decreasing due to aging, until it reaches a predicted state 130 equivalent to the minimum state 17 before the deadline for the predicted maintenance 7. In the illustrated case, the possible maintenance decision is indeed an insufficient one: it does not allow a predicted state 130 higher than the minimum state 17 until the end of the period 70. visionary. The intersection of the curve with the minimum state 17 corresponds to the pump failure deadline 19. According to one embodiment, for maintenance 6 to be performed without a maintenance decision identified as sufficient, and for at least one insufficient decision given for said maintenance 6 to be performed, the end-of-life deadline for equipment 2 is determined to be said failure deadline 19 associated with said maintenance decision. In other words, if, for a maintenance 6 to be performed, all possible maintenance decisions prove insufficient (to allow for zero failures during and until the end of the planned period 70, within the framework of scenario 8), then equipment 2 is at the end of its life and the maintenance 6 to be performed corresponds to the last maintenance in the life of equipment 2.When maintenance 6 is identified as the last maintenance in the life of equipment 2 and for a given maintenance decision, the end-of-life deadline for equipment 2 is then determined to be the failure deadline 19 associated with said maintenance decision.
[0167] When a planned maintenance 7 is identified as the last maintenance in the life of the equipment 2, the same procedure is followed to determine the end-of-life deadline of the equipment 2 associated with a given maintenance decision and the following scenario 81. According to an embodiment method and for maintenance 6 to be carried out (or for planned maintenance 7) identified as the last maintenance in the life of equipment 2, we select from the possible maintenance decisions the one that optimizes any combination between said failure deadline 19, a last use margin (MDU) and the constraints of tasks 90 of said last maintenance decision. In other words, for the maintenance 6 to be carried out, identified as the last maintenance in the life of equipment 2, all possible maintenance decisions are considered. For each of these maintenance decisions, the desired minimum operating time (MIND) for equipment 2 during the forecast period 70 is considered, namely the minimum operating time determined according to the constraints of the tasks 90 relating to said maintenance decision (i.e. according to the nature and volume of said tasks 90).
[0168] For example, consider the case of a truck engine-type piece of equipment 2, for which five possible maintenance decisions have been identified for the last maintenance in the life of the equipment 2, and for each of which the desired minimum operating time (MINT) has been determined, namely: - the El decision which replaces the diesel filter (with a desired minimum operating time DM1N of 3 days); - decision E2 which replaces injectors and diesel filter (with a desired minimum operating time DM1N of 1.1 months); - decision E3 which replaces injection pump and injectors (with a desired minimum operating time DM1N of 2 months); - Decision E4 which replaces the injection pump, injectors and diesel filter (with a desired minimum operating time DM1N of 2.1 months); - Decision E5, which replaces the injection pump, injectors, diesel filter, and cylinder rings (with a desired minimum operating time DM1N of 12 months). Among the possible maintenance decisions, each decision is then selected according to whether the life extension (PDV) associated with said maintenance decision (i.e., the time between said last maintenance in the life of equipment 2 and the end-of-life deadline of equipment 2) is compatible with the desired minimum operating time (DMIN) taking into account said maintenance decision (i.e., PDV > DMIN).
[0169] For example, we consider the aforementioned case of the truck engine with its five possible maintenance decisions identified for the last maintenance to be carried out in the engine's life, for each of which the desired minimum operating time (MIND) and the life extension (LDE) have been determined, namely: - the El decision which replaces the diesel filter (with a minimum operating time DM1N desired of 3 days and a PDV extension of 1 month, the injectors then being the limiting factor); - decision E2 which replaces injectors and diesel filter (with a desired minimum operating time DM1N of 1.1 months and a PDV extension of 1.5 months, the injection pump then being the limiting factor); - the E3 decision which replaces the injection pump and injectors (with a minimum operating time of DM1N of 2 months desired and a PDV extension of life of 0.5 months, the diesel filter then being the limiting factor); - the E4 decision which replaces the injection pump, injectors and diesel filter (with a minimum operating time DM1N desired of 2.1 months and a PDV extension of 6 months, segmentation then being the limiting factor); - Decision E5 which replaces the injection pump, injectors, diesel filter and cylinder segmentation (with a minimum operating time DM1N desired of 12 months and a PDV extension of 7 months, the crankshaft then being the limiting factor).
[0170] Decision E3 does not allow for a PDV lifespan extension (0.5 months) compatible with the desired minimum DM1N operating time (2 months). The same applies to decision E5. The selected decisions are therefore the Decisions E1, E2, and E4, each allowing, conversely, a PDV (Delivery Life) extension exceeding the desired minimum operating time DM1N. For each possible maintenance decision thus selected for the final maintenance in the equipment's life, the invention determines, for each predictable controllable usage condition 110 of scenario 8: - the last use limit, defined as the maximum possible limit, compatible with zero failures during and until the end of the desired minimum operating time DM1N of the equipment 2, after said maintenance to be carried out. The invention calculates this last use limit in a manner similar to the maximum limits 18. Said last use limits thus limit the usage limit to be respected during the forecast period 70 for zero failures until the end of the desired minimum operating time DM1N of the equipment 2; - the gap 182, for said condition 110 forecast of controllable use, between the target value in scenario 8 and the value of the last use limit.
[0171] For each possible maintenance decision selected, the margin of last use (MDU) is thus deduced for all the 110 predictable controllable usage conditions of scenario 8, as being the average of said deviations (i.e. similarly to the calculation of the margin of use before failure (MUAP)).
[0172] For example, we consider the aforementioned case of the truck engine, with its three possible maintenance decisions (E1, E2 and E4) selected for the last maintenance in the equipment's life and for each of which the margin of last use (MDU) has been calculated, namely: - the El decision which replaces the diesel filter (with an MDU margin of 3%); - Decision E2 which replaces injectors and diesel filter (with an MDU margin of 4%); - Decision E4, which replaces the injection pump, injectors, and diesel filter (with a 12% MDU margin). The optimal final maintenance decision is then determined from among the previously selected final maintenance decisions, including, but not limited to, the following: - which optimizes the extension of lifespan (PDV); - or which optimizes the last usage margin (MDU); - or which optimizes the last maintenance yield (LML) defined as the difference between said life extension PDV and said minimum operating time DM1N, said difference being referred to said life extension PDV (i.e. LML = 1 - DM1N / PDV); - or which optimizes any other criterion, combining, at the operator's discretion, at least two of the criteria between the extension of service life (PDV), the margin of last use (MDU), the yield of last maintenance (RDM); - or, failing a maintenance decision that satisfies the aforementioned criteria, the decision to replace equipment 2 with new equipment.
[0173] For example, consider the aforementioned case of the truck engine with its three possible maintenance decisions (E1, E2, E4) selected for the last maintenance in the equipment's life 2 (to allow the condition PDV > DM1N), namely: - the El decision which replaces the diesel filter (with a PDV extension of 1 month, an RDM yield of 90% and an MDU margin of 3%); - the E2 decision which replaces injectors and diesel filter (with a PDV extension of 1.5 months, an RDM yield of 26% and an MDU margin of 4%); - Decision E4 which replaces the injection pump, injectors and diesel filter (with a PDV extension of 6 months, an RDM efficiency of 65% and an MDU margin of 12%).
[0174] Furthermore, in the example, we choose to select the optimal last maintenance decision, according to the criterion that optimizes both the life extension (LD), the margin of last use (MDU) and the last maintenance yield (LML). Decisions E1 and E2 make no operational sense (the extended lifespan of the device is only 1 and 1.5 months, respectively). Furthermore, these decisions offer insufficient user comfort before failure (with a maximum service life margin of 3% and 4%). Conversely, decision E4 makes much more operational sense, with a 6-month extended lifespan, acceptable user comfort before failure (maximum service life margin > 10%), and a very acceptable return on investment (ROI) efficiency (65%). At the end of this final maintenance period in the engine's life, it is therefore rational not to replace the engine with a new one and to carry out the final maintenance decision E4 (i.e., the one including the replacement of the injection pump, injectors, and diesel fuel filter). The invention thus provides instructions for extending the engine's service life. For the maintenance 6 to be performed, identified as the final maintenance in the life of equipment 2, the invention thus selects, from among the possible maintenance decisions, the one that optimizes any combination between said service life extension PDV (which takes into account said failure date 19), the margin of last use (MDU), and the final maintenance efficiency RDM (which takes into account the duration DM1N and, in this respect, the constraints of the tasks 90 of said final maintenance).The invention thus provides instructions for extending the service life of equipment 2 and thus determines the optimal service life of equipment 2. It should be noted that this instruction for extending the service life of the equipment takes into account both the manufacturing and maintenance history 9 and . from history 11 of use of equipment 2, as from scenario 8 of use of equipment 2 as part of its life extension.
[0175] The same procedure can be used to determine the optimal final maintenance decision for any planned maintenance 7 identified as the final maintenance in the life of the equipment 2 and associated with the following scenario 81. According to one embodiment, the invention determines the optimal manufacturing decision 4 of the equipment 2, namely the critical manufacturing tasks 90 that optimize the predicted state 130 of the equipment 2 at the end of the predicted period 70 and for a scenario 8. For this purpose and for at least two fictitious equipment of the same series of said equipment 2, associated with distinct manufacturing decisions - a simulation is carried out by submitting said model 16 to said at least two manufacturing decisions and at least one predicted usage scenario 8; - model 16 generates at least one forecast state 130 for each of the two fictitious equipment; - one is selected from the aforementioned at least two manufacturing decisions based on the projected state 130 of the aforementioned two fictitious equipment; - The selected manufacturing decision is accessible to a designer / manufacturer. In other words, the invention considers the different possible manufacturing decisions 4 (i.e., the different possible combinations of critical tasks 90 for said manufacturing 4). Each combination is associated with a fictitious piece of equipment from the series.
[0176] For example, consider the case of a pump-type piece of equipment 2 for which the critical tasks 90 of manufacturing 4 are reduced to the following tasks 90: - "installation and selection of the pump bearing", with two possible bearing type options: types A and B; - "Installation and selection of the pump impeller", with three possible impeller type options: types A, B and C. The different possible combinations of critical tasks 90 of manufacturing 4 (i.e., the different possible decisions of manufacturing 4) are therefore the six combinations [bearing type; impeller type]: [A;A] [A;B] [A;C] [B;A] [B;B] [B;C]. We then consider the different hypothetical equipment, each associated with a manufacturing decision 4 from among the different possible combinations of manufacturing 4, assumed to be located in installation 3, assumed to have operated during an initial operating period 5 (reduced to a single operating period, i.e., without prior maintenance 10) within the framework of a usage history 11 (identical for all hypothetical equipment considered). We also consider a maintenance decision for maintenance 6 to be carried out (identical for all hypothetical equipment considered), a scenario 8 (identical for all hypothetical equipment considered). For each fictitious piece of equipment, the invention then uses said model 16 as a simulator, individualized to said fictitious piece of equipment, to determine the predicted state 130 of said fictitious piece of equipment at the end of the predicted period 70, by adapting the arguments submitted as input to model 16, namely: - the manufacturing and maintenance history 9 (reduced to the manufacturing decision 4, specific to the fictitious equipment audit); - the maintenance decision under maintenance 6 to be carried out (common to all fictitious equipment considered); - the usage history 11 (common to all the fictitious equipment considered); - the scenario 8 (common to all the fictitious equipment considered).
[0177] Thus, for all the hypothetical equipment associated with their respective manufacturing decision 4 (and all other parameters held equal), the projected state 130 of said hypothetical equipment at the end of the projected period 70 is compared. Having compared these projected states 130, the best projected state 130 designates the best manufacturing decision 4 (best combination of critical manufacturing tasks 90).
[0178] According to one embodiment, the simulation by model 16 determines the optimal manufacturing decision 4, namely the tasks 90 that optimize the optimal life cycle of equipment 2 (the optimal life cycle including the sequence of optimal maintenance decisions for the maintenance 6 to be performed and for each of the maintenance 7,71 forecasted in the life of equipment 2, the sequence of limit uses (i.e., the maximum limits 18) for each of the forecast periods 70,710 in the life of equipment 2, as well as the optimal lifespan of equipment 2), and this for a given usage profile (set of successive scenarios 8,81 for the different forecast periods 70,710 in the life of equipment 2). To do this, for at least two fictitious pieces of equipment from the same series of said equipment 2, associated with two distinct manufacturing decisions 4: - we perform recurring simulations in a similar way for each of the said at least two manufacturing decisions 4, to determine, a series of optimal maintenance decisions, a series of maximum limits 18 as well as the optimal lifespan of the fictitious equipment; - we choose the optimal manufacturing decision 4 based on the results of said simulations; - The selected optimal manufacturing decision 4 is accessible to the designer / manufacturer. In other words, since model 16 can simulate the predicted state 130 of a fictitious piece of equipment from the same series, as well as in the case of the actual equipment 2 (as seen previously), the invention determines the optimal life cycle of said fictitious equipment, as well as in the case of the actual equipment 2. Preferably, one This involves performing as many optimal lifecycle simulations as there are hypothetical pieces of equipment (i.e., as many simulations as there are possible manufacturing decisions). The optimal manufacturing decision is then selected based on the results of this simulation, that is, the one that allows for the best optimal lifecycle. System 1 thus makes it possible to determine, for the benefit of the manufacturer, the manufacturing decision that optimizes the complete lifecycle of the equipment in the series. In another embodiment, model 1 also makes it possible to quantify the sensitivity of the equipment's behavior (in particular the predicted states and the optimal lifecycle) to the operating conditions and to identify the optimal value for each operating condition: this indicates as many possible ways to improve the equipment's protection under the operating conditions, or even to optimize the equipment's operating point.
[0179] According to one embodiment, the invention provides an operator with overall supervision of all the equipment 2 in the series that it operates, installed in its fleet of facilities 3 (for example, a fleet of vehicles). To this end, the system 1 is applied to a fleet of several pieces of equipment 2 in said series at the same operator's premises. The results obtained for each of said pieces of equipment 2 are combined, and then said results are accessible to at least said operator. In other words, system 1 provides a given operator with, and updates at any time, information relating to the optimal lifecycle for each piece of equipment 2 in its fleet, for one or more desired usage profiles (namely, all the scenarios 8,81 of short- and long-term use over the lifespan of each piece of equipment 2), namely: - the program of optimal maintenance to come for each piece of equipment 2 (i.e. the sequence of optimal maintenance decisions, under the maintenance 6 to be carried out and all the maintenance 7,71 planned over the lifetime of each piece of equipment 2); - the continuation of the limit uses that will be possible to make of each piece of equipment 2 and over the long term of its life (i.e. for each period 70,710 forecast of operation in the life of each piece of equipment 2 and within the framework of scenarios 8,81); - the optimal lifespan of each piece of equipment. 2. System 1 then allows for the most accurate anticipation and updating at any time: - for the benefit of the actors in the "supply-chain", the nature of the parts needed for each of the maintenance operations to come in the life of each piece of equipment 2 (i.e. for the maintenance 6 to be carried out and all the maintenance operations 7,71 planned to come in the lifetime of each piece of equipment 2); - for the benefit of the maintenance provider, the scheduling of industrial maintenance (nature of tasks 90 to be carried out for maintenance 6 to be carried out and all maintenance 7.71 planned to come in the lifetime of each piece of equipment 2); - for the benefit of the operator, all the constraints (including the total cost) of future maintenance in the life of each piece of equipment 2 (for the maintenance 6 to be carried out and for all the planned maintenance 7.71 in the life of each piece of equipment 2); - for the benefit of the operator, the programming of the possible future use of each piece of equipment 2, over the long term of its lifespan, taking into account the sequence of previously determined usage limits. Thus, system 1 allows in particular: - to evaluate and update at any time during the life of each piece of equipment 2, the residual value of the equipment 2; - to determine the target price of the asset constituted by each piece of equipment 2 and thus to make a relevant decision to keep or resell the equipment 2.0n. Consider the example of a fleet of equipment 2 of the truck type. System 1 allows the operator of said truck fleet to answer the following questions: - What is the maximum possible use that the operator can make of several new trucks, between two successive maintenance periods, for different operating profiles (for example, in terms of load to be transported, mileage to be covered, average speed, average gradient of the roads used, ambient air temperature), associated with different geographical regions (for example, such as Russia, Senegal, France)? The operator owns several trucks that have reached mid-life and wishes to dispose of them. Considering their respective histories (9 and 11), and their operating profile specific to the region in which they are used, what is the remaining potential of each truck (in terms of load capacity, mileage, average speed, and lifespan), and what is the total cost of anticipated maintenance (7) over the remainder of their lifespan? At what price can the operator negotiate for each truck? - The operator owns several trucks previously used in Senegal that are nearing the end of their service life. Considering the typical usage profile specific to Senegal and the specific profile in France, and taking into account their respective usage and maintenance history, is it better to continue operating the trucks in Senegal or to repatriate them to France? - For each truck in the fleet, at what point does it become unprofitable to operate the truck in question, considering its historical data (9 and 11), its optimal end-of-life date, the maximum possible use until the end of its life, and the total cost Any upcoming maintenance? The invention allows for the updating, at the desired intervals, or even in real time, of the optimal lifecycle information specific to each truck in the fleet to answer these questions: the invention thus enables optimized and timely decisions regarding the management of these truck assets. System 1 therefore allows for the updating, at any time, or even in real time, and at the desired intervals, of the optimal lifecycle information for each piece of equipment 2 in the operator's fleet: this allows for optimized and proactive decisions, to the benefit of both the operator and the maintenance provider.
[0180] The invention makes it possible to gain significantly in control of the design, manufacture, maintenance and use of the equipment 2 of a given series, with long-term visibility of the life of each piece of equipment 2. The invention therefore allows the following gains in terms of control, safety and ease of use and economic gains, to the benefit of the designer, the manufacturer, the maintainer and the operators.
[0181] The invention unlocks access to control of optimal operating and maintenance decisions for a given equipment 2, when these decisions are short-term (i.e. concerning maintenance 6 to be carried out and use during the forecast period 70). On the one hand, the invention makes it possible to determine the optimal maintenance decision for the maintenance 6 to be carried out: this decision is individualized to the complete history of the equipment 2 (manufacturing and maintenance history 9 and usage history 11), is adapted to the desired use of the equipment 2 (scenario 8) after said maintenance to allow zero failure until the due date of the following planned maintenance 7, while pushing back, as much as desired, the limits of use compatible with the desired zero failure over the planned period 70. Furthermore, the invention eliminates the uncertainty of failure in a controlled manner. During operation, the invention provides the operator with real-time updated information on the maximum usage limits, consistent with zero downtime until the next maintenance: the operator now knows the limits to which they can push the use of equipment 2 without compromising the desired zero downtime, or within which they must restrict usage to achieve the desired zero downtime. The invention thus unlocks access to the safety and ease of use of equipment 2 and installation 3. Indeed, by eliminating the uncertainty of failure during operation in a controlled manner, the invention disrupts existing paradigms. The operator's question is no longer whether they will experience a failure or whether their failure prediction is correct. It is now the operator who decides not only the timing of the failure (for example, at the due date of the next scheduled maintenance 7), but also the operating margin before failure (i.e., the margin between the desired usage in scenario 8 and the maximum usage compatible with zero failure): it is now the operator who controls the use of equipment 2 to keep it compatible with the desired zero failure. Thus, the invention provides a genuine solution that meets the high availability requirements of equipment 2 (or the safety requirements of installation 3, when the availability of equipment 2 is a prerequisite for the safety of installation 3). Short-term economic benefits therefore arise from the invention. At the maintenance stage 6 to be carried out as during the operation of the equipment 2 during the planned period 70, the invention cancels the uncertainty of failure in operation concerning the equipment 2 and in a controlled manner (by access to the optimal maintenance decision and then by updated knowledge of the limit use compatible with zero-failure).
[0182] On the one hand, the invention thus enables the full-time production of the installation 3 desired by the operator. On the other hand, the invention provides superior reliability guarantees in terms of maintenance and operation: the invention therefore optimizes insurance contracts for maintenance providers and operators. The invention also unlocks access to control over long-term operational and maintenance decisions. Indeed, in terms of long-term use of equipment 2, the invention makes it possible to determine, as accurately as possible and at any time, the potential of equipment 2, by anticipating the uses that can be made of said equipment 2 over the long term of the remainder of its life (thanks to updated information on the limit uses of equipment 2 and its optimal life). In terms of long-term maintenance of equipment 2, the invention also makes it possible to anticipate, as accurately as possible, at any time, and this over the long term of the life of equipment 2 (and for each possible use profile of the equipment): the industrial tasks for the different maintenance to come, the total cost of the maintenance to come, the necessary and sufficient stocks of spare parts (and thus relaxes, for supply chains, the constraint in reactivity in the face of the unforeseen need for spare parts). Furthermore, the invention allows, at any time in the life of equipment 2, for the evaluation and updating of its maximum remaining potential as well as the total cost of the remaining planned maintenance 7 in the life of equipment 2, the invention also allows for the determination, at any time in the life of equipment 2, of the residual value of the equipment, as well as for relevantly informing the decision to keep or resell the equipment. Updating optimal lifecycle information for a piece of equipment or a fleet of equipment at the desired frequency, or even in real time, enables optimized and timely decision-making regarding asset management. The invention also allows for better control over the design and manufacturing choices of the equipment in the series, optimizing the designer / manufacturer's market strategy and providing superior reliability guarantees in design and manufacturing. This increased control is accompanied by economic benefits through the optimization of the designer / manufacturer's insurance contracts. Furthermore, the invention allows for the optimization of the designer / manufacturer's market strategy (market positioning). For a usage profile associated with a market segment for which equipment 2 is intended, the invention makes it possible to characterize the optimal life cycle of equipment 2 by generating associated metrics (margin between target and limit uses, optimal lifespan, optimal manufacturing cost and maintenance program cost over the lifespan). These metrics allow for quantifying the suitability of equipment 2 for the segment in question. The invention therefore enables the designer / manufacturer to: - a gain in terms of market strategy: the designer / manufacturer is then able to target more priority segments on which equipment 2 optimizes the suitability to the segment; - A gain in terms of market positioning: the designer / manufacturer is then better able to demonstrate the suitability of equipment 2 to the needs of the segment to customers and competitors, supported by model metrics. Furthermore, the invention also provides superior reliability guarantees in design and manufacturing. The invention makes it possible to determine the optimal manufacturing parameters adapted to a given usage profile (for example, adapted to a given market segment combined with a given geographic segment). The invention also makes it possible to identify potential areas for design optimization for a given usage profile. Thus, the invention allows for the optimization of the designer / manufacturer's insurance contracts.For this reason, the invention is disruptive, particularly its model 16, compared to conventional machine learning projects, by systematizing the integration of functional expertise and data and improving the consideration of functional expertise in data processing. In doing so, the invention is disruptive, especially compared to the state of the art, particularly in light of predictive maintenance technologies, as it is more relevant, more powerful, and offers broader applications than simple failure prediction.
Claims
Claims
1. Digital system (1) for supervising the operation and maintenance of at least one piece of industrial equipment (2) within an installation (3), executed by at least one computer terminal, comprising at least the following steps: - an installation is carried out within an installation (3) of at least one piece of industrial equipment (2) resulting from a manufacturing (4) and representative of a series, then said equipment (2) is at least operated within the framework of a period (5) until a maintenance step (6) to be carried out; - at least one forecast maintenance (7) is defined after said maintenance (6) to be carried out, following at least one scenario (8) with forecast conditions (110) of use of said equipment (2) over a forecast period (70) of operation, characterized in that the manufacture (4), installation, operation and maintenance of said equipment (2) induce at least: - a manufacturing and maintenance history (9) including: tasks (90) of manufacturing said at least one piece of equipment (2) up to said installation, possibly tasks (90) of at least one maintenance (10) prior to said equipment (2); - a history (11) of use of said equipment (2) on said period (5) between the installation and said maintenance (6) to be carried out, said history (11) of use comprising conditions (110) of use of said equipment (2) during said period (5); - a history (120) of the state (13) of said equipment (2), said state history (12) comprising hardware indicators (120) of said equipment (2); in that by means of a technical analysis of said equipment (2), at least one correlation (14) is determined between, on the one hand, at least one of said tasks (90) and / or at least one of said conditions (110) of use and, on the other hand, at least one of the material indicators (120) of said state (13), said correlation (14) establishing at least one link between causes of aging and consequences of aging of the equipment (2); in correlation (14): - the tasks (90) are characterized by those identified as critical, and / or the conditions (110) of use by those to which the equipment (2) is sensitive and exposed during operation or when stopped, the tasks (90) and conditions (110) considered impacting the state (13) of said equipment (2); - the material state (13) of the equipment (2) is characterized by the indicators (120) identified as being representative of this state (13) of said equipment (2); then, in correlation (14), we determine: - measured physical quantities or functions of the measured physical quantities characterizing the conditions (110) of use to which said equipment (2) is sensitive and exposed in operation or when stopped; - measured physical quantities or functions of measured physical quantities characterizing the material state (13) of the equipment (2) at a given time; Then for the other equipment of said series, data associated with these tasks (90) are recovered and extracted, and data associated with the physical quantities or functions of physical quantities relating to these conditions (110) of use and to these material indicators (120), as identified in said correlation (14), so as to obtain a set (15) of data; training of at least one virtual model (16) is carried out, on the basis of the data set (15); and in that during the maintenance (6) to be carried out on said equipment (2), the following values are submitted to said model (16): - on the one hand, at least one of the tasks (90) of the manufacturing and maintenance history (9) and at least one of the conditions (110) of use of the use history (11), - on the other hand, at least one of the tasks (90) of the maintenance (6) to be carried out and of said forecast conditions (110) of use of the scenario (8); said model (16) generating a forecast state (130) of said equipment (2) subsequent to said maintenance (6) to be carried out, said forecast state (130) being compared to a minimal state (17) identified as required for the operation of said equipment (2).
2. Supervision system (1) according to claim 1, characterized in that - at least one variation is made of at least one of the values of the tasks (90) of the maintenance (6) to be carried out: - when the values are submitted to said model (16), the values of said variation are introduced; - among all the variations, at least one sufficient decision of the maintenance (6) to be carried out is selected, for the forecast state (130) of said equipment (2) greater than or equivalent to the minimum state (17), at the time of said forecast maintenance (7).
3. Supervision system (1) according to one of claims 1 or 2, characterized in that - for a given maintenance decision, a modification is made to the value of at least one of the forecast conditions (110) of use of the scenario (8); - when said values are submitted to said model (16), the values of said modification are introduced as well as the values of said maintenance decision; - a limit is calculated for at least one of said forecast conditions (110) for the forecast state (130) of said equipment (2) equivalent to the minimum state (17), at the time of the forecast maintenance (7).
4. Supervision system (1) according to claims 2 and 3, characterized in that - when the values are submitted to said model (16), the selected values of said sufficient maintenance decision are introduced; - a maximum limit (18) of said forecast condition (110) is calculated for this sufficient maintenance decision, for the forecast state (130) of said equipment (2) equivalent to the minimum state (17), at the time of the forecast maintenance (7).
5. Supervision system (1) according to claim 4, characterized in that - a margin of use is determined for at least one of the forecast conditions (110) of use of said scenario (8), as being the difference (182) between the corresponding value and the corresponding maximum limit (18).
6. Supervision system (1) according to claim 5, characterized in that - an optimal decision is selected from among the sufficient decisions, as having the acceptable margin of use or as having at least one of said acceptable deviations (182).
7. Supervision system (1) according to claim 6, characterized in that it comprises - a graphical representation in the form of an abacus, with at least one curve associated with at least a first of the forecast conditions (110) as a function of a second of said forecast conditions (110), said abacus determining the maximum limit (18) of a first forecast condition (110).
8. Supervision system (1) according to any one of the preceding claims, characterized in that - in the correlation (14), the manufacturing and maintenance history (9) is reduced to a history of critical tasks (90) in the form of at least one list of successive values, each of the values in the list characterizing the task (90) in question at a given maintenance, in each list, only the remanent value is chosen as being the value adopted at the last maintenance at which the task (90) in question was carried out - only the remanent values are kept in the history (9) of critical tasks (90).
9. Supervision system (1) according to any one of the preceding claims, characterized in that in the correlation (14), the functions of the measured physical quantities of the conditions (110) of use comprise - a calculation of the time of presence of the measured physical quantities in at least one range of values; and / or - a calculation representative of at least one fluctuation of the measured physical quantities: and / or - a counting of said at least one fluctuation.
10. Supervision system (1) according to any one of the preceding claims, characterized in that in the correlation (14), at periodicity, - the recovery of new data is repeated from at least one manufacturer, a maintainer and / or an operator, - then the extraction of said new data to obtain a completed data set (15), - followed by updating the learning of said model (16) on the basis of said completed data set (15).
11. Supervision system (1) according to any one of the preceding claims, characterized in that after said maintenance (6) once carried out, values are submitted to said model (16): - on the one hand, of at least one of the tasks (90) of the manufacturing and maintenance history (9), of at least one task (90) of the maintenance (6) carried out and at least one of the conditions (110) of use of the history (11) of use since said maintenance (6) carried out - on the other hand, of the said forecast conditions (11) of use of the scenario (8), said model (16) updating the forecast state (130) of said equipment (2), said forecast state (130) being compared to a minimal state (17) identified as required for the operation of said equipment (2).
12. Supervision system (1) according to claims 5 to 11, characterized in that we carry out at least the following steps: - it is assumed that at least said sufficient decision of the maintenance (6) to be carried out has been carried out and it is assumed that the scenario (8) has been executed up to the forecast maintenance (7) following said maintenance (6) to be carried out; - then at least one variation of at least one of the values of the tasks (90) of said forecast maintenance (7) is carried out; - when the values are submitted to said model (16), the values of said variation are introduced as well as the values of a following scenario (81) planned for the forecast period (710) of operation following said forecast maintenance (7); - among all the variations, at least one sufficient decision of said forecast maintenance (7) is selected, for the forecast state (130) of said equipment (2) higher than or equivalent to the minimum state (17), at the time of the maintenance (71) following said forecast maintenance (7); - at least one of the maximum limits (18) is determined as well as the margin of use associated with said sufficient maintenance decision thus selected and said following scenario (81); then said steps are repeated in a recurring manner for any other subsequent predictive maintenance in the life of the equipment (2).
13. Supervision system (1) according to claims 6 and 12, characterized in that - said optimal decision is selected from at least said sufficient decision for the corresponding maintenances.
14. Supervision system (1) according to claims 2 to 13, characterized in that - when a maintenance decision is insufficient with a forecast state (130) lower than said minimum state (17), the failure deadline (19) is determined for said corresponding maintenance decision.
15. Supervision system (1) according to claim 14, characterized in that - for the maintenance (6) to be carried out or for a forecast maintenance (7, 71) without a maintenance decision identified as sufficient, and for at least one given insufficient decision of said maintenance (6) to be carried out, the end of life deadline of the equipment (2) is determined as being said failure deadline (19) associated with said maintenance decision.
16. Supervision system (1) according to claim 15, characterized in that - for the maintenance (5) to be carried out or for a predictive maintenance identified as the last maintenance in the life of the equipment (2), the one which optimizes any combination between said breakdown deadline (19), a margin of last use and the constraints of the tasks (90) of said last maintenance is selected from among the possible maintenance decisions.
17. Supervision system (1) according to claim 1, characterized in that for at least two fictitious equipment of the same series of said equipment (2), associated with distinct manufacturing decisions - a simulation is carried out by submitting to said model (16) said at least two manufacturing decisions and at least one forecast usage scenario (8) over the assumed lifetime of said two fictitious equipment; - the model (16) generates at least one forecast state (130) for each of said two fictitious equipment; - one is selected from said at least two manufacturing decisions based on the forecast state (130) of said two fictitious equipment; - the selected manufacturing decision is accessible to a designer / manufacturer.
18. Supervision system (1) according to any one of claims 12 to 17, characterized in that - for at least two fictitious equipment of the same series of said equipment (2), associated with two distinct manufacturing decisions; - recurring simulations are carried out in a similar manner for each of said at least two manufacturing decisions, to determine the optimal life cycle associated with each of said manufacturing decisions: a series of optimal maintenance decisions, a series of maximum limits (18) and margin of use associated with these optimal maintenance decisions as well as the associated optimal lifetime of the fictitious equipment; - the optimal manufacturing decision is chosen according to the results of said simulations; - the selected optimal manufacturing decision is accessible to said designer / manufacturer.
19. Supervision system (1) according to any one of the preceding claims, characterized in that - it is applied to a fleet of several pieces of equipment (2) of said series at the same operator; - the results obtained for each of said pieces of equipment (2) are combined; - said results are accessible at least to said operator.
20. Supervision system (1) according to any one of the preceding claims, characterized in that - the learning of said model (16) falls within the field of artificial intelligence and may be machine learning.
