Augmented reality applied to design for disassembly assessment for a volumetric pump with rotating cylinder

ABSTRACT Design for Disassembly (DfD) and Augmented Reality (AR) have become promising approaches to improve sustainability, by providing efficient delivery and learning assets. This study combines DfD and AR to deliver a method that helps to streamline maintenance processes and operator training. It focuses on a common part in the process industry that requires frequent maintenance and repair. DfD was applied to the pump’s design to ease disassembly and reduce material waste, energy consumption, and maintenance time. AR was used to provide an interactive guide to improve the operator understanding of its internal parts and assembly/disassembly procedures. The resulting DfD-AR led to a reduction in maintenance time and shows potential to deliver better training. This highlights the potential of DfD and AR to enhance sustainability, learning, and productivity. The resulting disassembly sequence was taken to an AR simulation, helping process designers to better understand the procedure and further optimize the solution with other constraints.


Introduction: design for disassembly and augmented reality
Design for disassembly (DfD) is a method that allows to find the disassembly sequence of a varying mechanical complexity (Abuzied et al., 2020;Rasmussen et al., 2019).The purpose of DfD is to obtain a sequence list of all the operations (tasks) to perform, taking account with related information (timing, methods, tools, and operators) needed to dismantle all the components of the product or only those necessary to extract a certain component (target).This is particularly useful for those operators dedicated to maintenance processes, who can follow a clear work schedule for complete or necessary disassembly of the parts (Kissi et al., 2019).It is clear how this approach provides advantages in the maintenance of industrial machinery formed by hundreds of components, particularly in terms of time, costs, and resources.The maintenance and service that is performed on the components can be preventive or corrective.In the first case, the element is replaced in advance, before a possible failure.In the second case, the replacement or repair of a damaged component happens after the fact (Manfè et al., 2013).Therefore, DfD allows the straightforward servicing of a component, a portion of it, or the recovery of the material and its reuse.DfD is also important for environmental impact, since it reduces the consumption of materials and allowing reusage of components, as shown in Figure 1, decreasing the pollution generated.
In conclusion, DfD helps to develop the product life cycle by incorporating the recovery of parts or materials at the end of the life cycle, it improves the quality of processing, it offers more flexibility, and it reduces disassembly time and costs, consumption of resources, and waste (O'grady et al., 2021).An essential aspect of the DfD is the 'disassembly depth', which is the total number of disassembly steps to perform.The removing total cost curve of a component has an approximately linear trend with the depth of disassembly, unlike the revenue.Specifically, we are talking about the Disassembly Sequence Planning (DSP): a special application of DFD for the disassembly sequence planning of the product.The DSP aims to reduce the complexity of disassembly to get to the specific component that requires servicing.For this reason, it is crucial to consider some important guidelines: the choice of recyclable materials, the reduction, standardization, and simplification of fasteners and connecting elements (O'grady et al., 2021).In addition, it is necessary to consider the movements of the operators considering their fatigue and effort, the handling of work tools, the structure of the product, the removal sides of the components, the various dimensions, ergonomically and operational constraints.There are two kinds of disassembly processes: a complete disassembly when all elements are removed, or a selective disassembly, when only the necessary components are removed to get to a particular component (target).Disassembly can also be destructive or non-destructive, depending on the necessity or not to break a component to remove it.The following methods that will be presented offer non-destructive disassembly.However, researchers have found several challenges in DSP, as a different approach proved to be optimal according to the part complexity and scope of the sequence obtained, varying those results from a selective disassembly sequence planning (SDSP) for repair and maintenance; differentiated from a partial DSP (for precious material recovery) delivered by Anil Kumar et al. (Anil Kumar et al., 2021); until a complete DSP (for safe disposal at end-of-life products) as outlined by Gunji et al. (Gunji et al., 2021).Artificial Intelligence nowadays helped researchers to portray combined approaches, based on an optimal function gathering to reduce disassembly times by considering tool changes, gripper changes and directional changes as parameters.A case study on a hybrid, Exponential Moving Average (EMA) -with Particle Swarm Optimization (PSO) was proposed by Gulivindala et al. (Gulivindala et al., 2021) and compared the calculations made by both instances.
Moreover, other research approaches on product disassembly are ruled out by the development of various mathematical models to better calculate the most energyefficient, viable sequence by means of a stochastic disassembly network plot, combined with a number of considered disassembly decision-making properties (G.Tian et al., 2012); another proposal (G.Tian et al., 2022) handles a new heuristic method with multiple targets to minimize smoothness, as well as disassembly-related costs and energy consumption; as well as a low-carbon inspired design disassembly line balancing fruit fly optimization algorithm (MDFOA) (Yang et al., 2019); lastly, an approach to find the optimal, energy-efficient scheduling problem (Z.Tian et al., 2023) with a number of constraints of diverse resources tuned up to a small-batch, dynamic flexible job shop for the aerospace industry.
Furthermore, Disassembly Geometric feasibility (GF) is ultimately needed in order to identify potential collision paths in robotic assembly path design (Kumar et al., 2022).The gathering of a feasible assembly sequence to be performed by a robotic arm also represents utter positive assembly costs that are important to industries to develop a cost-effective manufacturing process (Raju Bahubalendruni & Biswal, 2015).Researchers have aid to develop an algorithm that can successfully solve an assembly sequence generation with account of the exploded view creation altogether (M.V. A. R. M. V. A. Bahubalendruni et al., 2019).Likewise, an assembly sequence that uses more parallel possible stable subassemblies proved to reduce the overall assembly time in large-scale production facilities (M.V. A. Bahubalendruni et al., 2019).
Consequently, Augmented Reality (AR) technology will be applied in this study: a highly interactive visual method used to present digital information in the physical (real) space (Chang et al., 2017;Xiong et al., 2021).AR is used in many industries: medical, education, marketing, military, video games and manufacturing.This technology can be implemented through devices equipped with a camera and AR tracking and rendering software that recognizes useful images, shapes, objects, points, and curves.The device sets the virtual elements in defined points located in the space, to reproduce them in the real space.AR instructions can be easily published and displayed on a smartphone or tablet, and in many other types of company devices.In the manufacturing industry, AR offers a better way to create work instructions that are easily accessible, by overlapping digital content to real working environments.Augmented reality also enhances the frontline workforce by improving the dissemination of information (Harikrishnan et al., 2021;Jing et al., 2021;Radu et al., 2021) resulting in faster transfer of know-how, modernized training methods (Fernández-garcía, 2021), and instant remote access to skills, and ultimately higher customer experience.Augmented reality is redefining the way frontline employees acquire information and how they can digitally interact with the physical space around them (Li et al., 2021;Masclet et al., 2021).It is transforming human processes by accelerating skills development and supporting orientation (Hassan et al., 2021;Sampaio & Almeida, 2018;Singh et al., 2021).This transformation results into faster execution, fewer manual processes, and the ability to make better decisions (Martins et al., 2021).Industrial organizations in a variety of sectors around the world are already realizing how valuable a company-wide AR implementation is from an economic standpoint as well, by improving business results to the point of achieving the fastest and most significant ROI (Return of Investment), even towards achieving sustainability at work (Joerss et al., 2021).The first industrial companies to adopt augmented reality solutions are saving millions in operating costs by improving the productivity of frontline workers and reducing scrap, rework, and waste.AR offers benefits in several areas of practical application: can be leveraged to improve production, technical support (Bolano et al., 2021), and sales (Qin et al., 2021), thus in marketing (Tan et al., 2022).For example, Toyota has improved safety and communication in its factories by connecting subcontractors and employees remotely (Langlotz & Aurich, 2021), eliminating the need for traveling in person to the site to solve problems.Volvo Group simplified its quality control processes by converting digital information into quality control operating instructions (Rosales et al., 2021).These instructions allow technicians to examine engines quickly and accurately, as they leave the production line.Lastly, in a different industry, Cannondale improved its service process for independent bicycle repair shops to adapt 3D CAD data into interactive repair manuals (Auffinger, 2021): these manuals allow technicians to quickly identify the parts to be replaced without the need to perform lengthy research through instruction manuals or catalogs to order replacement parts.
The aim of this study is to portray a solution for a Disassembly Sequence for a regularly produced element that takes into account maintainability, reuse of the parts and constraint of movements by means of the following two theoretical methods of sequence calculation.The obtained sequence will be verified in a real case-inspired, Augmented Reality application, where this information can be exploited to obtain further information in benefit of the product design, production and after-sales areas.

First method
The first method studied for the research of the best disassembly sequence proposed by Mandolini et al. (Mandolini et al., 2018), based on the previous study of (Yi et al., 2008) and has become a feasible approach for assessing DfD (Alrufaifi, 2021).It is based on the concept of disassembly wave propagation: we choose the target component to be worked, then we imagine that a wave starts from this component and expands through all the components adjacent (Figure 2).A first wave, at the time instant τ1, crosses all the elements adjacent to the target component; then, at the time instant τ2, a second wave will cross the adjacent components, and the process repeats on each element.We repeat the procedure until all the components of the machine are crossed.Possible disassembly sequences are generated by removing all components that are present in the various waves, starting from the outermost (at time τn) to the innermost (τ1) until the target component.An important rule to remember is that it may be necessary to remove more than one component belonging to the same wave level, so during the process several removal steps can be performed (Su et al., 2021).

Second method
The second method studied is based on the construction of a disassembly precedence matrix (DPM) (Laili et al., 2022), and a procedure proposed by Fei Tao et al. (Tao et al., 2018), which contains the information about the overall dimensions and spatial constraints and offers partial or parallel disassembly sequences.The DPM is a square matrix of n rows and n columns, considering the n components that form the mechanical product, and is formed in turn by four sub-matrices: (1) The FFM (fasteners-fasteners matrix); (2) The CFM (components-fasteners matrix); (3) The FCM (fasteners-components matrix); (4) The CCM (components-components matrix).
Moreover, it is important to indicate the removal directions of each component, through a sequence [-X, +X, -Y, +Y, -Z, +Z] that indicates the removal directions of the components for each vector of the three Cartesian directions (X, Y, Z).Examining the drawing of the product, we will consider the X-axis directed along the pump axis, the Y-axis upwards and the Z-axis directed out of the sheet and towards the observer.The sequence will consist of six digits of only 0 and 1, writing 1 at the component removal direction and 0 for the other directions (Table 1).
In the DPM matrix, every element contains the succession [-X, +X, -Y, -Z, +Z] of 0 and 1: the CFM submatrix, at the i-th row reports the limitations of the i-th component due to the presence of the j-th fastener in its removal direction.So, if a certain j-th fastener hinders an i-th component along its removal direction, in the corresponding box will be inserted 1, otherwise 0. The same is true for the other sub-matrices.Whenever a component is removed, the corresponding row and column will be deleted, so the matrix is dynamic and decreases as disassembly process proceeds.It is possible to remove a component if a 0 appears at least in one Cartesian direction in the respective row of the DPM matrix.The method allows partial or parallel disassembly, or disassembly in modules, disassembly of blocks of several components and subsequent disassembly in parallel.We show the rules to follow for the creation of the disassembly sequences: -R1: removal of the j-th fastener (Fj) if FCM(j,:) = 0 and FCM(j,:) = 0.
A third rule is added to these two rules in case of module disassembly, considering that within a module the relationships between the parts represent internal limitations to it, thus they do not affect the separation between modules.
-R3: module disassembly if no 1 appear in the rows of the DPM matrix, except for internal relations.So, to generate a partial or parallel disassembly sequence, five steps must be followed: -P1: find any sequence of fasteners; -P2: disassemble the fasteners in the order found.finding a sequence for the removal of a target component but offers the possibility to find all disassembly sequences (partial or complete) that, if properly compared, can offer an optimal sequence for a target component that you choose.
A comparison between the two methods is analyzed in Figure 3.

Table of disassembly weights: time calculation
To measure the duration of the disassembly sequences found, a methodology present in research from Mandolini et al. (Mandolini et al., 2018), that allows to find out total Time Measurement Units (TMU) used by each strategy.Other outcomes of Deng-Zhi, Chen, et al. (Chen et al., 2020) optimized the time method through a flow balance in the AND/OR network for a large-scale disassembly evaluation.This method has the flexibility to optimize such time-calculus by generating disassembly subsets based on disassembly precedence relationships.In addition to the proposal of a selective parallel disassembly sequence planning (SPDSP) by Zhang, Lei, et al. (Zhang et al., 2021).However, this process establishes a summary table containing the estimates of the disassembly times of each type of component, calculated considering appropriate parameters that we will show.For each parameter, the relative weight is shown, i.e. disassembly time, expressed in TMU (Time Measurement Unit), i.e. a specific unit of measure used for these procedures.The conversion into minutes or seconds is immediate: 1 (TMU) = 0.0036 (seconds).Proceed as follows: evaluate the weight of each parameter (force, weight, position . ..) of a component and then multiply them and obtain the corresponding value in TMU, repeat these steps for all components, add up all the values in TMU obtained and finally the final value is converted into seconds.For example, if the shaft component is considered, a weight is assigned to each of its parameters; these are multiplied to obtain the resulting value expressed in TMU.The procedure is iterated for all the components of the sequence, all the values obtained in TMU are added together and converted this value in seconds: Tot.Seconds = (Tot.TMU) * 0.036 (sec).
This mathematical procedure based on the use of tables is particularly advantageous in terms of time, compared to a practical measurement in the workplace, implemented by timing every single activity of the operators, because it allows knowing in advance, almost immediately, the duration of an activity.It is therefore easy to understand how this method facilitates the comparison between disassembly sequences during the design phase.

Augmented reality disassembly sequence simulation verification
Improvements on digital imaging technology developed during the last few years has given the possibility of improving the quality of artificial environments, by adding detail to the ambient created.High-quality environments have enabled users to portray realistic environments using fully digital platforms.
Therefore, Augmented reality (AR) is emerging as next-generation interactive element able to provide realistic three-dimensional (3D) visual experiences (Xiong et al., 2021).This technology has proven to be valuable in diverse fields such as education (Iatsyshyn et al., 2020), healthcare (Viglialoro et al., 2021), engineering (Hansen et al., 2021;Noghabaei et al., 2020) and gaming (Lehto et al., 2020), among many other areas.AR fosters the interaction between user, and the realistic ambient through digital contents, by portraying virtual images whilst maintaining a see-through capability.Nevertheless, AR has proved to be a viable alternative to enhance the understanding of production and manufacturing processes (Mourtzis et al., 2019), aiding the user to better understand the matters of a regular production environment (Lai et al., 2020), as well as quality control processes (Szajna et al., 2020).

Overview of piston pumps
Piston pumps are the best choice for high-pressure applications or those requiring great performances.They belong to the class of variable displacement pumps and offer a long service life and great efficiency.Radial piston pumps (Figure 4) are very robust radial devices with high volumetric efficiency, wide speed range and high reliability, even in severe conditions.The pump studied has a closed pump body, an intake line and a delivery line, is equipped with a hollow piston, with a spring inside that moves it, and radial to a shaft in which it is inserted.RN 3 × 11.8 type roller bearings are inserted between an inner and outer ring.The two side shims act as containment.Given the simplicity of the device and the small size, there are no bearings between the shaft and the body: for heavier tasks, the use of bushings or rolling bearings would be appropriate.

Cad modelling
Inventor, a 3D programming software produced by Autodesk, was used for the modelling of the product.It was chosen to use this program because it is simple, intuitive, practical, immediate and with a vast library of components, which proved to be very useful in this case.The first thing to do is to start by designing each component separately, following     the Bill of Materials (BOM) of the product that is shown below in Table 2 to then assemble them.At a first glance of the drawing, it is easy to see how this product has different types of connections: ▪ shaft-hub connections between the shaft and the body, between the piston and the shaft; ▪ elastic connection between spring and piston; ▪ threaded connections given by the screws and the nut; ▪ connection by key.Elements such as the body, the shaft, the piston, the distributor, the delivery and suction fittings have been designed starting from the realization of the 2D sketches, respecting the dimensions detected by the drawings shown in Figures 5 and 6, and then using 3D model development functions such as extrusion or revolution depending on the components.The bearing (components 3,4 and 5) and the sealing ring were downloaded from the SKF website, a Swedish Company founded in 1907, operating in the mechanical sector, manufacturer of rolling bearings, sealing elements, lubrication systems, of which it offers downloadable 3D models.The two shims and the ring (component 6) were downloaded from the Inventor content center, a real library of components present in the software.The spring, pinion, screws, washers and tab elements were created in the design section, identifying the specific command for each component, as described below.The spring is one of the fundamental components of this device, since it guarantees the movement of the piston, but it is also one of the most stressed components and therefore subject to wear.It is a cylindrical, compression helix spring subjected to an axial load, which, acting at the ends, can cause an increase or decrease in the pitch of the helix.The spring material under load is mainly stressed in torsion.For its design, the appropriate compression spring generator command was used, in which it was sufficient to enter appropriate design parameters and then have the system doing the calculation.The pinion was processed through the spur gear generator command, another command found in the component design section.It was also necessary for this to insert appropriate calculation parameters relating to the gears.The screws and washers were made with a command generating bolted connections, dimensioning them on the existing holes and making the relative thread.The key connection generator was used for the key, entering the dimensions of the key and proceeding with the creation of the relative grooves on the shaft and pinion.Finally, only after having created all the components we did move on to the construction phase: we started from the body (component 1), to then import one component at a time and assemble it respecting constraints, joints and degrees of freedom, up to the assembly final (Figures 5, 6 and 7).

First method application
The purpose of the first method is to find the best possible sequence to remove a certain component: the spring has been chosen as the objective component because, as previously mentioned, it represents one of the components most subject to stress and wear.Therefore, more likely its removal might be required, for maintenance or replacement.
We apply this method by reporting below the construction of the disassembly wave diagram (Figure 8): elements 3, 4 and 5 are considered as a single component, the bearing, both for simplicity in writing the program and in operation, since the removal of these components occurs in a single process.
The sequence obtained is as follows: 19-12-13-15-18-14-11-17-6-2-3-16-7-8-10-9.Remember that removing components 14 and 15 means removing three washers and three screws respectively, but for simplicity in the representation of the disassembly wave graph they are shown as single elements.However, observing the product component detail drawing it is possible to avoid removing the components present in the third disassembly wave τ3 of the diagram: the sealing ring (7), the key ( 16), the two shims (2), the ring (6) and bearing (3).Their permanence in the mechanism does not significantly affect the process since it is sufficient to apply a little extra force to remove the shaft (8).Furthermore, also the components 12 and 13 cannot be removed, since they do not represent an impediment for the removal of the component 11, much less for the rest of the disassembly.Therefore, the sequence can be simplified as follows: 15-14-11-19-18-17-8-10-9.

Second method application
Before proceeding with the application of the second method, the same considerations made for the previous method are assumed regarding the choice of the target component (spring) and the bearing component 3 is renamed, as the set of elements 3,4 and 5.
It is important to identify the ways of removing the components, as explained previously in the presentation of the method, also identifying the fasteners.For the body component, the direction of removal is not indicated because it is considered fixed for the purposes of disassembly, all components are extracted from it (Table 3).
Subsequently, the precedence matrix is written (disassembly precedence matrix or DPM): for clarity, each sub-matrix is reported first and then followed by the entire matrix (Table 4).Also, in the following matrices the body component does not appear, for the same reason explained above.The element consideration is summarized in Figure 9.
Then proceed by following the steps shown previously in the explanation of the method, first removing the appropriate fasteners, and then moving on to the next components: remember that it is not necessary to remove all the fasteners before proceeding with the other components.The sequence found by the analysis of the DPM matrix is: F15-C14-C11-F19-C18-C17 -C8-C10-C9.
Observing this last sequence obtained, it can be noted that the disassembly is uncomplete and components F12 and F13 had not been removed; in fact in this second method, after removing component C11, two modules M1 = (C11-C12-C13) and M2 are formed; they are composed by the remaining elements and, although the method foresees the parallel disassembly of the modules, in this case it is not necessary to dismantle the module M1, since the spring is present in the second module M2.

Comparison of results and calculation of disassembly times
Both methods led to the same result, even though for the first case, the use of practical sense, together with the knowledge of the product component layout, allowed to exclude certain components from the disassembly process.The common result was obtained thanks to the simplicity of the product component, which allows disassembling mainly along a single direction, so there are few disassembly alternatives.However, between the two methods, the second one is what provided the optimal solution directly during its application, as it is known that the second method foresees additional complexity characteristics of the product, as summarized on Table 5 while the first served mainly to identify all possible sequences with a more simple approach.The optimal sequence is therefore the following: 15-14-11-19-18-17-8-10-9.The next step is the evaluation of the disassembly times relatives to the best performing sequence found, using the disassembly weight tables presented previously.Thereafter, in the construction of the table of disassembly weights of the pump components, the values of the body have not been reported since it is considered a fixed element (Table 5).
To conclude the discussion on Design for Disassembly, as already mentioned, for the evaluation of the best disassembly sequence other equally important factors must be taken into account in to the disassembly time.It is in fact essential to make an economic evaluation considering the workforce costs as well as the necessary tools and to include the cost related to the safety of the operators and the working environment, which involves the use of dangerous and harmful substances such as lubricants, oils and various residues.

AR application: unity and vuforia
An augmented reality application capable of showing a product being disassembled can play a key role in streamlining its maintenance process.Instead of equipping maintenance workers with instruction booklets and drawings difficult to understand, it makes it possible to show them how to proceed with the removal of various components.The order they will follow will be the optimal one to get to the end of the process using little time and the least number of resources (technological tools, machinery, number of workers, and so on).Both the industrial process and the learning of the workforce thus becomes immediate.This represents a factor entirely to the benefit of business productivity.The application in augmented reality here proposed shows how the costs to be faced by a company are rather small, but the potential that comes from the simplicity of applying the method is high.At the hardware level, in fact, it is sufficient to equip oneself with a few augmented reality visors.
In this case study, the augmented reality software Unity and Vuforia are used to create an explosion of the rotating cylinder pump assembly in augmented reality.Vuforia is a software development kit (SDK) of augmented reality, a sort of library that must be integrated into Unity.It provides fast and accurate image tracking allowing the virtual content to be superimposed on the images of the surrounding environment.First, a new project is created in Unity.Vuforia's online portal is used to create a license for the AR camera in Unity, characterized by an auto-generated code.Subsequently, an image target is developed: this is the object in the real world that the software recognizes when the camera records it, thanks to a set of geometrical target points located in the imagine areas with more particulars ().A very simple picture was created, and its final evaluation was 5/5 on Vuforia, which is the best.The Vuforia package was imported into Unity, then the AR camera, into which the license code and the target image were copied.It was necessary to convert the component assembly file into obj format and then load it into Unity and insert it as a child of the target image.This obj format file was recognized by the program, but it is monochrome: therefore, it was necessary to proceed with coloring each component separately using materials texture and realistic colors (Figure 10).
Finally, individual animations of the assembly were created, representing the exploded object, following the simulation of both disassembly sequences found previously and comparing them.Each component was added to the animation and its position within the timeline was determined.
For an AR representation, the marker is framed with the webcam, so the device will automatically display the 3D assembly during the animation phase (Figures 11 to 14).
Moreover, further use of having this 3D assembly would allow to create individual scenes for every disassembly step, as well as indicate the correct direction of the parts dismantling.Figures 14 allow to see the Augmented reality video for servicing the pump.This animation would allow practitioners to be trained correctly to the handling operations of this type of mechanical element.Finally, it is appropriate to underline the many other possibilities that can be proposed by having shown the features of the proper 3D assembly treatment possible by Unity, it is easier to imagine how this technology is appropriate and valuable in a business setting, across most of its R&D, and helping thru product production, logistics, and aftersales departments, as  main information for procedures, and training of worker force afterwards, who must handle complex devices, hence facilitating their work and learning the processes overall.

Discussion
This case study demonstrated that an AR solution to assist the studies of disassembly design methods and technologies that enable product servicing, maintenance processes of parts and machines, as well as to better understand product manufacturing procedures otherwise just stated mathematically.
From an economic and productivity perspective, design by focusing on disassembly deliver methods that redefine the manufacturing and production techniques to optimize the processing, working and safety conditions, as well as reducing time, costs and overall resources needed.
Accurate servicing attention is also important for environmental sustainability: It allows for the recovery and recycling of components or materials, to reduce consumption, waste, and pollution.The technology of AR is a guide to understanding the processes and activities of industrialization, thus improving the productivity.

Conclusions
• The effectiveness of the design for disassembly applied to a small product such as the rotating cylinder has been demonstrated.• This methodology is particularly useful and convenient in a series production context with hundreds of components and machines to be checked.Its convenience for the management of resources and from an economic point of view have been verified.• A correct sequence assessment would allow to considerably reduce resources and direct and indirect dismantling costs.• The advantages can also be evident in the speed of intervention, having the immediate availability of the necessary processing sequences and times.• The methodology of Augmented Reality assessment, used in numerous sectors in addition to the industrial one, has been applied to ease the virtual verification of the theoretical process portrayed by two sequence methods calculated in this exercise.• The methodologies used in this study are not much widespread among companies yet, as resources and necessary skills that are not guaranteed for the use of this tools.

Figure 1 .
Figure 1.Resources reprocess to develop circular economy.

Figure 2 .
Figure 2. Example scheme for the first method.(a) Schematization of the positioning of components in the assembly; (b) Analysis of all components close to the target at time instant 1; (c) Continuation of the analysis for affected components at the next time instant.

Figure 3 .
Figure 3.Comparison between the two methods.

Figure 4 .
Figure 4. Representation of the constructed product.

Figure 5 .
Figure 5.View of the construction in half section on Inventor.

Figure 6 .
Figure 6.Representation of the final product on Inventor.

Figure 7 .
Figure 7. Exploded view of the rotary cylinder pump on Inventor.

Figure 8 .
Figure 8. Wave scheme applied to the mechanical pump.

Figure 9 .
Figure 9. Flow chart for the second method.

Figure 10 .
Figure 10.Assembly in Unity with the image target.

Figure 13 .
Figure 13.Augmented reality video: fully assembled pump for servicing.
the final sequence has been found.If not, proceed to the next step; -P5: check for modules that satisfy rule R3, if present disassemble them in parallel, otherwise go back to step P2.This second method, unlike the first, is not aimed at