Introduction

Current researchers and practitioners are extremely fascinated by the adoption of information and communication technology (ICT) as it will be highly beneficial (Hollenstein 2004) to all fields of business. Analysis of composite project management can be efficiently done with the help of project management services (Stoshikj et al. 2014). Existing IT service quality measurement framework is refined for the IT service industry (Lepmets et al. 2014). In continuation of this, the field of education is significantly improved due to the adoption of ICT. Recent trend in education is to implement different types of educational programs using ICT (Bidarian and Davoudi 2011). While peeping into the world of literature, a plethora of research can be witnessed in internet based control systems accessing data wirelessly from a remotely located process. Monitoring and controlling of such systems are done by using wireless and internet communications technologies (Thompson 2004).The rising reliance of businesses over information and information usage with ICT tools and mediums, have formed an inevitable requirement of information security in organizations (Singh et al. 2013). Some multiplicity aspects together with geographic remoteness and cultural differences impact the success of ICT projects (Bohm 2013).

A wireless communication system is a flexible system and has a number of advantages; besides the mobility of the devices within the environment. Adding a communication device to the system or removing one from it without any disruption to the remainder of the system is cakewalk. Keeping apart the initial outlay on setting up the cell sites, the cost of running and maintaining a radio based communication system is minimal. These coupled with other factors prove the demand for a radio communication system in the office environment. Being a public frequency network, wireless interface is extremely risky for official private information. As the number of users utilizing the same frequency increase, the rapidity and the feasibility of the wireless signals decline. The shortcomings of the aforesaid wireless network can be overcome to a great extent by using high speed point-to-point wireless networking and wireless backhaul. Point-to-point wireless networking uses series topology instead of mesh, which increases the radial distance coverage and removes geographical barriers. This network, being an invisible one is not easily hindered and an alternate solution helps to restore the connectivity in the event of any interruption. Wireless backhaul is an economic connection, which can handle a huge amount of data uploading, and downloading in a blink of an eye (http://freewimaxinfo.com/point-to-point-wireless-networks.html, 15th January 2014).

Only a few researches have been reported in emerging systematic design methods using this technology or in designing ICT enabled process control systems. Several open loop wireless control systems such as wireless transmitter and ICT enabled control valves are available in the process industry. Nevertheless, owing to security issues, industrialists have not dared to implement these innovative methods for a closed loop system till date even on a trial basis. This provides the motivation for this study.

The objectives of the study are

  • To identify the elements of the enablers of wireless control system for oil refinery.

  • To develop total interpretive structural model for the enablers of wireless control system for the case organization.

Interpretive Structural Modeling (ISM) and Total ISM (TISM) Methodologies

The interrelationships among the various elements and their degree of association can be effectively assessed by ISM. ISM is used to translate vague and inadequately articulated rational models of systems into visible, well-defined models (Sushil 2012). It is used to determine the most prominent knowledge sharing barriers that deserve serious consideration (Sharma et al. 2012). Based on correlation, an overall configuration is hauling out from the complex set of variables (Mandal and Deshmukh 1994; Thakkar et al. 2007; Warfield 1974a, b; Faizal et al. 2006). ISM is effectively used to determine the critical success factors of R&D performance in Indian manufacturing firms (Tripathy et al. 2013). In an ISM, the interpretation of the digraph is done at two levels, i.e., at the nodes and at the links. Though an ISM interprets the nodes by defining the elements representing it, the interpretation of links provided is comparatively weak. The links simply indicate the contextual relationship between the elements in terms of the direction of relationship in a paired comparison. The causal thinking behind the directed link in terms of how it operates remains implicit as it is neither captured nor represented in the ISM.

In the ISM of enablers of wireless control system, the contextual relationship between the elements is clarified. However, the interpretation and logic behind the interrelationship between the elements of enablers are not elucidated. Most important enablers can be selected only by knowing the exact meaning of the relationship between elements. An ISM interprets the association in terms of the contextual relationship between each pair of elements, and the direction of their relationship. There is a requirement to interpret the relationship further in terms of causal thinking behind the associations which can be accomplished with the help of TISM. In TISM technique, the limitation of ISM is tackled by using the tool of interpretive matrix (Sushil 2005a) where the causal thinking is captured at the time of collecting data from the experts and is made explicit in the matrix. The model case study reported in the present work details the application of TISM model for the enablers in the implementation of a flexible control of plant using ICT. After the model is developed, it is further subjected to assessment by a different group of domain experts so as to enhance its validity.

For the development of TISM model, brainstorming sessions were conducted with the top and senior level managers of an oil refinery where the case study was conducted and these suggestions were analyzed to identify the elements of enablers. The TISM (Sushil 2012) is a pioneering extension of ISM that is used to develop a hierarchical structure of the set of variables of interest. The TISM developed a directed graph for the complex system amongst the set of variables, which helps transform poorly articulated mental models into a systematic form that can be used for many interpretations. This is an innovative qualitative modeling technique that has been used by researchers in various fields of investigation (Nasim 2011; Prasad and Suri 2011; Wasuja et al. 2012; Srivastava and Sushil 2013; Dubey and Ali 2014).

As in the case of ISM, reachability matrix and its partitions are adopted to develop the TISM model. The steps of the basic process of TISM are briefly outlined as follows.

  1. Step I

    Identify and define elements. The first step in TISM modeling is to identify and define the elements whose relationships are to be modeled. The elements for modeling in the study are enablers related to the implementation of wireless control in industry and are identified through interviews with experts in an oil refinery.

  2. Step II

    Define contextual relationship. To develop the configuration, it is essential to define the contextual relationship between the variables of interest. Here, the contextual relationship identified between the elements of enablers is “enabler A will influence or enhance enabler B”. An example may be “enhanced efficiency will influence or enhance ease of maintenance”.

  3. Step III

    Interpretation of relationship. Traditional ISM seeks an interpretation of the relationship between the variables of interest, but remains almost silent to interpret how that relationship really works. However, in TISM, clarification of the interpretation is also available. In the context of this study, the interpretation will be “In what way enabler A will influence or enhance enabler B?” (Table 1). It will help achieve in-depth knowledge.

  4. Step IV

    Interpretive logic of pair-wise comparison. An “Interpretive Logic - Knowledge Base” can be created for pair-wise comparison of the elements of enablers. In this, the ith element is compared independently to all the elements from (i − 1)th to the nth element. If there are n elements, there will be all n(n − 1)/2 paired comparisons. Since each pair of elements (i, j) may have two possible directional links (i − j) or (j − i), there will be in all n(n − 1) rows in the knowledge base. For each (i − j) link the entry code could be “Yes”(Y) or “No” (N) and if it is “Yes”, then further interpretation is necessary. The interpretive logic-knowledge base for enablers is illustrated in Appendix I (Table 6). The total number of pair-wise comparisons is 11 * 10/2 = 55 and total number of rows is 11 * 10 = 110 for the knowledge base (Sandbhor and Botre 2014).

    Table 1 Elements, contextual relationship and interpretation for the enablers of wireless control system in process industry
    Full size table
  5. Step V

    Reachability matrix and transitivity check. The reachability matrix for the enterprise perspective is prepared and the entries 1 for Y and 0 for N from the interpretive logic-knowledge base. The matrix is checked for the transitivity rule (if A–B and B–C, then A–C). For each transitive link, the knowledge base is restructured as Y, and in the interpretation column, ‘transitive’ is entered. If there is a significant interpretation available, it is written along with the ‘transitive’ entry.

  6. Step VI

    Level partition on reachability matrix. The level partition is carried out, similar to ISM, to understand the placement of elements level-wise (Warfield 1974b; Saxena et al. 2006). The reachability, antecedent and intersection sets for all the elements are determined and arranged in a table. The elements in the top level of the hierarchy will not reach any elements above their own level. As a result, the reachability set for a top level element will consist of the element itself; and any other element within the same level which the element may reach whereas, the antecedent set consists of a group of enablers itself and other enablers, which may help achieving it. Therefore, intersection of the reachability set and the antecedent set will be the same as the reachability set if the element is at the top level. The top level elements satisfying the above condition should be removed from the element set and the same procedure is repeated iteratively till all the levels are determined. These levels are exploited for the formation of digraph and the TISM model.

  7. Step VII

    Developing digraph. The elements are arranged graphically in levels and the directed links are drawn as per the relationships shown in the reachability matrix. Digraph or the directed graph, illustrates the relationship between the elements of enablers as per the numbers assigned to them. Relationship between elements j and i can be represented by an arrow which points from i to j. A simpler version of the initial digraph is obtained by eliminating the transitive relationships step-by-step by examining their interpretation from the knowledge base. Only those transitive relationships may be retained that have significant interpretation.

  8. Step VIII

    Interaction matrix. The final digraph is translated into a binary interaction matrix form representing all the interactions by ‘1’ entry. The cells with ‘1’ entry are interpreted by selecting the pertinent interpretation from the knowledge base in the form of interpretive matrix.

  9. Step IX

    Total interpretive structural model. The information obtained from the interaction matrix and digraph is used to derive the TISM. The nodes in the digraph are reinstated by the interpretation of elements placed in boxes. The interpretation in the cells of interpretive direct interaction matrix is portrayed by the side of the respective links in the structural model. This leads to the formation of total interpretation of the structural model.

About the Refinery and Case Study

This study was done with the help of an oil refinery situated in, central Kerala, India. It was established in 1963. As the oil refinery is reluctant to reveal its identity, in this paper it is being referred as ABC refinery. It has a production capacity of 9.5 million metric tons per annum. The ABC refinery has presently embarking on the integrated refinery expansion project which would increase the refining capacity from 9.5 to 15.5 million metric tons. Founded as a public sector company owned by the Government of India in 1963, it was acquired by a leading private sector company in 2006, which presently owns and runs it. Fuel products of ABC refinery includes liquefied petroleum gas, naphtha, motor spirit, kerosene, aviation turbine fuel, high speed diesel, and fuel oils. The ABC refinery entered into the international petroleum business stream in January 2001. Since then, the refinery has earned the reputation as a reliable player in the international deal and the refineries fuel oil has been benchmarked in the various foreign countries fuel oil markets. The refinery has been blessed with fine scenery and the entire complex, scattering across over thousand acres has been so constructed as to merge naturally with it.

Wired systems are used for monitoring and controlling the process, thus consuming more space. The plant officials have an idea to innovate the plant to get enough space, better control and efficiency, greater flexibility and optimum plant operation. When the authors approached plant officials with their proposal, they instantly arranged for a model case study as a prerequisite for the implementation of the proposal. TISM model was used for the study. The study proposes a systematic approach that presents judgment on the best implementation strategy. To implement the project, data on user perspectives have to be generated.

Conventional control system used in the ABC oil refinery is as shown in the Fig. 1. Long cables are used for connecting controllers located in the control room and sensors placed in the field. Many wires are essential for this. Therefore, these types of control systems are limited to a specified length of cable and are not flexible. The demerits of the wired systems can be eliminated by using a flexible wireless control system.

Fig. 1
figure 1

Cable based closed loop control (prior implementation of case study)

Full size image

The following points are of utmost importance for the optimal management of the field instrumentation:

  1. (i)

    Remote access to all the parameters of the installed devices,

  2. (ii)

    Remote calibration and test,

  3. (iii)

    Automatic recording of all test results as well as storage of all the configuration data.

Wireless devices can make all these functionalities available in control room, similarly to what plant operators are used to get with wired smart devices.

In the proposed wireless control model, controller and the plant to be connected via a two-way digital communication channel of limited bandwidth. As depicted in Fig. 2, sensors are monitoring the state of a process and occasionally forward the digital measurements (Sk) to a remote controller. Based on these measurements, the remote controller computes a control signal (Uk) and sends it to the actuator over the feedback channel. Upon recovering the control signal, the actuator applies a suitable action to adjust the plant state. Consistent communication occurs only if both measurement Sk and feedback control Uk are decoded within specified time limit defined by the controller.

Fig. 2
figure 2

Wireless closed loop control (post implementation of case study)

Full size image

Rationale Behind the Selection of Case Study

Conventional control systems are used in the ABC oil refinery. Many wires are essential for this and hence it requires so much space. Addition of instruments to the system necessitates excess length of cables and more space. With the introduction of a complete wireless control system, connecting cables between instruments can be eliminated. Any number of instruments can be added to or removed from the system. Flexibility and mobility of wireless control system are very high. Because of these advantages of wireless control systems, a case study is incorporated in the said oil refinery.

Development of the Total Interpretive Structural Model

As an initial step, the general awareness of the refinery was made available through interviews to the experts in the company. Later, data that is more relevant was collected through brainstorming in a small expert group from the industry. According to the experts in the refinery, while adopting ICT enabled wireless control system technology, many important factors also should be considered. This enabled to verbalize an initial draft of enablers for the proposed work. Five experts with rich experience and deep knowledge in the plant management were identified for the study. Enablers derived from the opinion of experts have been interrelated using TISM.

Enablers identified were as follows:

  1. (1)

    E1: free from the hassle of cables.

  2. (2)

    E2: work efficiency under harsh conditions.

  3. (3)

    E3: large scale deployment.

  4. (4)

    E4: ease of maintenance.

  5. (5)

    E5: low earth faults noise.

  6. (6)

    E6: user friendliness.

  7. (7)

    E7: fast installation and commissioning.

  8. (8)

    E8: flexibility.

  9. (9)

    E9: enhanced efficiency.

  10. (10)

    E10: suitable for areas where network is scarce.

  11. (11)

    E11: centralized monitoring.

Table 1 shows the elements, contextual relationship and interpretation for the elements of enablers of wireless control system in process industry.

The contextual relationship identified was “Enabler A will influence or enhance Enabler B.” The interpretive logic-knowledge base was developed. As there were 11 enablers, the total number of rows in the knowledge base was 11 * 10 = 110. The authors discussed the knowledge base with the same experts and hunted their help to fill in the knowledge base. These responses were further used to develop the reachability matrix. For any pair-wise comparison, if 60 % of the responses were Y, it was taken as Y otherwise as N. All the responses for Y were analyzed in terms of the interpretations given by the experts and a combined statement integrating all responses was developed. Appendix I (Table 6) shows the interpretive logic-knowledgebase of the study.

In the second sub-step, final reachability matrix is prepared. For this purpose, the concept of transitivity is introduced so that some of the cells of the initial reachability matrix are filled by inference. Transitivity can be explained with the following example: if element ‘i’ relates to element ‘j’ and element ‘j’ relates to element ‘k’, then transitivity implies element ‘i’ relates to element ‘k’. Transitivity is the basic assumption in ISM and is always used in this modeling approach (Farris and Sage 1975; Sushil 2005a, b). It also helps in maintaining the conceptual consistency. Reachability matrix and the final reachability matrix with transitive links are shown in Tables 2 and 3.

Table 2 Reachability matrix
Full size table
Table 3 Final reachability matrix with transitivity
Full size table

The final reachability matrix consists of some entries from the pair-wise comparisons and some inferred entries. Here, while framing the correlation between various enablers transitivity is also considered. The transitivity rule states that if a variable A leads to a variable B and if B leads to another variable C, then A leads to C (Pramod and Banwet 2010 ). From the final reachability matrix, the reachability and antecedents of each element of enablers are developed. Reachability set consists of a group of enablers itself and other enablers, which may help to realize, whereas, the antecedent set consists of a group of enablers itself and other enablers, which may help achieving it (Attri et al. 2013; Singh and Sushil 2013). The enablers that are common in reachability set and antecedent set are assigned at the intersection set. The elements for which reachability and intersection sets are the same occupy the topmost level group (level I group). These top level elements are removed from the set for the formation of the next table. This process is repeated till all levels of each element are found. These levels are exploited for the formation of diagraph. These subsequent iterations are shown in Appendix II (Tables 7, 8, 9, 10 and 11).

Table 4 Level matrix
Full size table

Digraph is the synonym for directed graph, which illustrates the relationship between the elements of enablers as per the numbers assigned to them. Relationship between elements j and i, can be represented by an arrow which points from i to j. Digraph with significant transitive links is shown in Fig. 3.

Fig. 3
figure 3

Digraph with significant transitive links

Full size image

The final diagraph is translated into a binary interaction matrix form Appendix III (Table 12). The information obtained from the interaction matrix and diagraph is used to derive the TISM (Fig. 4).

Fig. 4
figure 4

TISM model

Full size image

The hierarchical structure portrayed in Fig. 4 shows that the driving force related to the implementation of wireless control system in industry is “centralized monitoring”. “Free from the hassle of cables” is the driver, “fast installation and low earth fault noise” is the driving force. The lagging enablers are “user friendliness, flexibility, enhanced efficiency and large scale deployment”. This TISM structure clearly depicts a hierarchy with the lagging elements of enablers at the top. It is very comprehensible that for “large scale deployment” of the wireless system, the system must develop action plans for easy maintenance, to be ideal even when the network is inadequate and to be free from the disturbances of cables.

Validation of TISM

Even though a TISM model is more meaningful than an ISM, it is much more demanding in terms of time required from the experts. Experts not only need to define the contextual relationship among each pair of variables but are also expected to provide logic behind each response. For example, while developing the TISM model for enablers of the flexible control system, an expert had to define and provide reason for 110 [i.e., n(n − 1)] pairs of relationships.

‘Scarcity of time’ being a general constraint for experts across all fields, it is not easy to find enough experts to volunteer for this stage. Nevertheless, after the model is developed, it is easier to find experts to evaluate the pictographic representation (TISM) of relationships, especially since the number of valid relationships reduces. In the present study of enablers, the numbers of credible relationships have reduced from 110 to just 20. Thus, an evaluation of the developed TISM based model for the enablers is also introduced in this paper, as a possible solution to tackle this widespread limitation, facilitating validation of the model from a larger expert base.

Thus, for developing the initial TISM model for enablers, responses from five experts were taken. However, after the model was finally developed, feedbacks regarding the approval of the model, from eight other domain experts have been elicited using the appraisal stencil. Such an approach facilitated validation of the model from more number of domain experts, which comprised of top level officials of seven different process industries in Kerala. In this regard feedback from the eight domain experts were composed and the acceptability of the model was established on a five-point scale ranging from “strongly agree” to “strongly disagree” in relation to the linkages and reason(s) given by previous experts. Each link in the model is accepted if the average score for the model is above three and the entire model is assumed to be acceptable if the average of all the links beat the score of three (Table 5).

Table 5 Assessment of TISM model
Full size table

It is clear from the table that, all links, except link 15 were accepted and also the entire model is accepted. The rejected link is dropped from TISM model and the validated TISM model developed as shown in Fig. 5.

Fig. 5
figure 5

Validated TISM model

Full size image

Conclusion

The fields of management are swarming with innumerable models and elicitations of experts. ISM is well proven to utilize the experts’ knowledge and experience (Kumar et al. 2013). An ISM interprets the associations in terms of the contextual relationship between each pair of elements, and the direction of their relationship (Sushil 2009). However, TISM uses the implementation of an interpretive matrix, which captures the causal thinking of the experts. In this paper, the enablers for the successful implementation of a flexible control system in the industry are identified and exemplified the use of a qualitative technique (TISM) to model these for greater understanding of the interplay of these enablers. As a possible innovative step to verify and enhance the validity of the model obtained, a model assessment has been added in addition to the demonstration of using the TISM technique. In general, this methodology of the modeling technique is expected to be worthy to the researchers; and for the domain practitioners the resultant model provided insights too. In this work, 11 enablers of ICT enabled wireless control system were identified and considered for analysis. This research illustrates that the developed TISM model is acceptable. From this, the importance of implementing the flexible control systems in industrial plant is evident. Implementation of the proposed flexible control system has become a necessity for industries to survive rapid changes and fierce competition that characterizes the current global business environments. The developed TISM model and its validation done herein for the enablers of the flexible control system encourage broad scale adoption of the technology in remote monitoring and control of plants.