Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Using resources responsibly is a paradigm drawing more and more attention. The consideration of aspects of sustainability in customer’s buying decisions and federal legislation makes many enterprises reconsider their industrial practices. The provision of mechanical energy accounts for 64 % of the overall industrial energy consumption in Germany (Brischke 2010). Saving energy requires a responsible and demand-driven use which may save costs, prepare enterprises for future challenges and decrease the dependency on volatile energy prices (Ghoshray and Johnson 2010; Ratti et al. 2011). Various guidelines like energy management systems (DIN 16001:2009-08 (2009); ISO (International Organization for Standardization) 2011) or the use of key performance indicators (Bunse et al. 2011) have been developed to manage and control energy use within production. Furthermore, energy-efficiency is widely considered an important requirement within the planning process of new equipment. Integrated machines like machine tools or larger production equipment are more and more crafted to fulfil these requirements or possess particular energy saving modes. Still, many of these measures do not have the expected impact on small and medium enterprises (SME). In this environment, production equipment is not replaced every time the manufactured product changes. It is more likely to be modified in order to fit new product characteristics. Thus, the used equipment tends to have a longer life expectancy in SME and adjustments due to technical improvements do not occur as often as in larger enterprises. In addition, SME often do not possess the needed expertise within the field of energy efficiency. Finally, transparency regarding energy consumption on equipment level or even on component level is often not accessible. In a nutshell, SME often lack “initiation, market information, expertise and financial resources” (Bunse et al. 2011).

The research project AssiEff aims at the development of a mobile assistance system for energy-efficient production. Based on an in situ measurement of the individual components and the knowledge of the incoming orders, recommendations to improve the efficient use of the production equipment will be given. According to their individual information needs, decision makers, production engineers and machine operators are addressed differently. This approach aims at the involvement of the employees−from shop floor to decision level−to improve energy-efficiency within order-related production in SME.

The chapter is structured as follows. First we take a look at possible strategies at the shop floor level, these may include recommendations to exchange components or to operate the system more energy-efficiently. As these recommendations strongly rely on the actual conditions of system operation, the chapter then describes potential scenarios for order-related manufacturing. In the next section, the relevant stakeholders are introduced, who may influence energy-efficiency measures in a manufacturing context. The following sections focus on the design of the support system which fuses data from sensors reading e.g., energy meters, using metadata and formalized heuristics as well as planning information provided e.g., by an ERP system. Non-functional requirements with respect to the IT-system which are not addressed by the previous sections are then collected in order to derive a basic system architecture. As the system is specifically tailored towards the needs of SME, we show possible deployment scenarios such as temporary deployment and the provisioning of energy efficiency optimization as a service. We provide the reader with related work before we conclude the chapter.

2 Strategies to Support Energy Efficiency of Existing Production Facilities on the Shop Floor Level

Due to the growing interest in resource efficiency, energy aspects within the planning process of new production systems have moved into the centre of attention. Also considerable work and attention was spent to the development of new and energy efficient machine components like drives, pumps or fans.

Besides that, the optimization of existing production sites has by far not been exploited yet. The estimated saving potential in this field extends up to 50 % of today’s energy costs (Schmid 2004; European Commission 2003).

To address these aspects, the ongoing research project AssiEff develops an assistance system for order-related, energy-efficient production. The achievement of the overall goal of a more efficient use of energy requires a stronger involvement of the employees. The projected assistance system aims to support decision makers, production engineers and machine operators in SME by a transparent display of the actual energy use and the recommendation of potential action. The approach is based on two basic strategies to save energy:

2.1 Exchange of Components

The life cycle of a production system commonly covers several product generations as well as various demand patterns and changes in production programs. Usually the active components of a production system, e.g., actuators, drives and pumps are designed to cover all possible demand patterns and to fulfil all possible workload conditions.

As a result, existing production sites frequently contain over-sized and therefore inefficient components. Recent surveys found that currently deployed electrical drives realistically reach energy conversion efficiencies of about 20 % (European Commission 2003) and also suggest quite similar figures for pneumatic components (Radgen and Blaustein 2001). Manufacturers estimate potential savings using state-of-the-art technology as high as 82 % of energy costs for compressors and above 50 % for components like pumps, hoisting drives or fans (Siemens 2011).

Despite this enormous potential, high-energy-consumption components are seldom replaced with newer, more energy-efficient ones tailored to the actual workload patterns (Schmid 2004).

Unlike during the planning of new production systems, energy-efficiency audits are not yet widely applied in the industry, specifically not in SME. Transparency of information about the actual energy consumption on component level is often missing and therefore, existing components are often reused over two or more product generations as replacing them seems more costly.

So, compared to current industrial practice in small and medium enterprises, the deployment of components dimensioned according to their predicted energy consumption may save energy and thus significantly save costs over their life span.

In the context of this strategy, we aim to estimate the ecological and economic benefits of an exchange of components in case of the reuse of an existing production system. Based on operating data and in situ measurement of the relevant components, their operating modes and the predicted utilization, energy savings and their economic impact in terms of return on investment are calculated. Furthermore, this data serves as a baseline to compare the actual utilization during operation to predict worthwhile component exchanges. The planning assistant aims to support decision-making based on a model of the component-level classification of energy-efficiency. Therefore, it needs realistic, timely input data, energy consumption of each component in different operating modes, operating hours, energy costs for kWh and information about alternative state-of-the-art components. The latter is necessary to provide a benchmark regarding the actual efficiency and cost improvements.

2.2 Energy-Efficient System Operation

Energy-efficient system operation as a second strategy addresses saving potentials that can be realized by dynamically controlling the use phases of production systems. By “Turn-Off-Engineering” and alterations of the order sequences, a more efficient use of resources is expected as it is shown in Fig. 1.

Within the field of order-related manufacturing, equipment is frequently left in idle states, consuming almost the same amount of energy as in full-load conditions.

As nowadays, a paradigm shift in the management of equipment from capacity utilization to flexible and adaptable production has already occurred, the utilization and thus energy consumption of machine equipment largely depends on its underlying demand patterns. Figure 2 shows the typical characteristics of energy use within order-related manufacturing. In the context of this research, tests were carried out at an automated production line for the customer-specific assembly of USB sticks. Short periods of production are followed by non-productive phases. Within these non-productive phases, energy of about 500 Watts is consumed by infrastructural parts of the equipment like data processing and energy supply for sensors and components. While producing, energy consumption is at about 2.000 Watts.

Fig. 1
figure 1

Exchange of components, triggered by the consideration of operating points of different product generations. Source Schlund et al. (2011)

Full size image
Fig. 2
figure 2

Energy use at an automated production line for customer-specific assembly of USB sticks

Full size image

Energy-efficient system operation aims at reducing energy consumption by the use of short-time switch-offs−either of the entire equipment or some of its parts. Depending on the energy consumption and the length of non-productive operation, this may include the definition of additional machine states like stand-by or simmer modes. The following preconditions are therefore required:

  • order-related manufacturing (significant non-productive phases);

  • fast starting and stopping devices;

  • real-time information about operating modes;

  • real-time information about upcoming orders.

Intelligent and IT-supported fast deactivation of components will be referred to as “Turn-Off-Engineering” (Schlund et al. 2011). Energy-efficient system operation, which is supported by real-time information about potential switch-off situations, closes inefficient gaps within the order status by the partial shutdown of the equipment. For the aforementioned production line, Fig. 3 displays the savings potential of the following aspects of “Turn-Off-Engineering”:

Fig. 3
figure 3

Savings potential for energy-efficient system operation

Full size image
  1. 1.

    Avoidance of idle states before start of production;

  2. 2.

    Avoidance of idle states between orders;

  3. 3.

    Use of separately switchable devices (drives);

  4. 4.

    Demand specific switching on/off of separate components during production process.

  5. 5.

    Optimization of switching on/off;

  6. 6.

    Parallel time-delayed manufacturing of multiple orders.

The measures 1–5 refer to different options of “Turn-Off-Engineering”, which may be used combined. The parallel time-delayed manufacturing of multiple orders displays a change of the initial order sequence. Using the load curve, the arrows pinpoint the saving potential of each individual measure which is as high as 30 % for one single measure in the depicted example (measure 6).

3 Scenarios to Describe the Range of Use Cases

In order to estimate potential benefits of isolated or combined measures of the described strategies, scenarios are used. Scenarios are defined as a management tool for identifying a plausible future and a process for forward-looking analysis (Ahmed et al. 2010; Gelman 2010). For the design of mobile assistance for energy-efficient production, they are employed to provide a range of possible situations that may occur in order-related manufacturing of a (partially) automated production line.

To specify the possible situations for energy-saving measures, three main parameters are considered for each scenario:

  • order intervals (orders/time);

  • processing time (time/order);

  • order variation (standard deviation of orders/time).

Combining these three dimensions and parameter values for each of the characteristics, 18 plausible scenarios could be identified. They describe the entire range from highly utilized equipment with constant load to low utilization with high variation. Figure 4 illustrates the results. Implausible scenarios caused by incompatible combinations of order interval and processing time (e.g., order Intervals: 1.000 orders/h and processing time: 100 s) are marked with an “X”.

Fig. 4
figure 4

Classification of scenarios according to their input data. Source Schlund et al. (2011)

Full size image

Further parameters are either considered as largely constant like operating costs (energy costs) or may be derived from the main parameters above, like operating hours (in terms of the total hours of operation which is dependent on processing time and order intervals).

The scenarios form the basis for the decision models of the assistance system. For example, in the case of alteration of order sequences (e.g., bundling), they represent the full range of possible demand patterns. Based on the simulation of randomly created order distributions within those scenarios, decision support can be provided for a wide variety of SME. Currently an algorithm is created to bundle or withdraw orders based on energy saving constraints. Regarding the equipment, the needed time gaps between orders to change to an energy-efficient operating mode (or even to switch-off) provide the command variable.

4 Applications for the Most Relevant Stakeholders

Within the field of operation on the shop-floor-level there are basically two fundamental strategies to implement energy-efficiency measures. First of all, machine characteristics already consider “eco”-modes and adapt their behaviour to the actual needs. This approach is based on “intelligent” behaviour of the equipment to analyze and evaluate its operational parameters and provided data to react in an energy-efficient way. The other approach is based on intelligent behaviour of the relevant stakeholders. Actions within this field aim at the implementation of an “eco”-mode within their behaviour. Compared to the machine-oriented approaches, they rather focus on “soft” measures to provide transparency of energy use and influence a more careful use, e.g., in times of less utilization.

Obviously, these two approaches do not have to be employed separately. Depending on the use case and the possibilities of interference, fully automated equipment would be hard to adapt in the scope of the second approach. However, partially automated production lines with a lot of loosely connected energy-consuming devices may benefit significantly from including the work force and considering soft factors with the goal of saving energy. Furthermore, this approach needs possibilities to interfere into the individual machine controls to turn-off individual components, stable start-up processes and easy-to-change components, thus a flexible and versatile production system.

The AssiEff project follows this strategy. Based on a real-time display of energy use and the proposal of possible actions to be taken by the relevant stakeholders, it aims at a more considerate use of energy. Furthermore, by the recommendation of potential measures, the assistance system leaves it to the recipient to decide about what action is to be taken. To address the right person within the context of energy-efficiency, a stakeholder analysis has been carried out and basically six relevant groups of persons could be identified (Table 1).

Table 1 Relevant stakeholders and their typical questions within the decision process for energy-efficient measures
Full size table

Out of these six, the three stakeholders which most likely have an impact on influencing energy-efficiency were considered the following:

  • The decision maker;

  • The production engineer;

  • The machine operator.

These roles are directly addressed by the assistance system through individual templates, which consider the specific requirements in terms of information provision and recommendations for possible action.

5 Adoption Barriers and Non-Functional System Requirements

Besides functional requirements, such as the use of scenarios and the consideration of the various stakeholders described in the previous sections, several non-functional requirements have to be met for the realization of the mobile assistance system.

For easy adoption in SME, it is necessary to lower barriers regarding the operation of the system. For example, it is desirable that the system can also be hosted by a third party vendor in order to reduce operation costs and efforts. For the same reason, it should also be possible to outsource operation and support of end user devices.

There are a number of barriers to the adoption of energy-efficiency projects. The barriers are quite heterogeneous, and vary from sector to sector (Kentor 2003; Schleich 2009). However, it is possible to generalize that in the commercial and services sectors, energy costs are usually below 3 % of the overall costs, and thus projects aiming at increasing energy efficiency within such organizations are often rejected simply because they are not seen as “strategic” (Kentor 2003). We address this problem by designing a system that enables outsourcing of large parts of the energy optimization process, reducing both the initial investment necessary to establish the process, and the necessary strategic commitment. Offering intermediate services through vertical service providers is a well-known strategy for supporting SME (Albino and Kühtz 2004).

Requirement I: The system should allow outsourcing of the metering and analysis processes to a specialized service provider.

As a result, it could improve the willingness of enterprises to adopt energy efficiency projects if the system was able to not only supply them with strategic information they can use when e.g., planning future purchases of manufacturing equipment, but also offers a way to identify immediate benefits (Lockett NJ and Brown 2005) that could be reaped by e.g., rescheduling tasks.

Requirement II: The system should allow for operational short-term analyses to enable immediate benefits for enterprises.

Alcántara et al. (2010) identify promising horizontal technologies for improving energy efficiency. The proposed technologies of metering energy and steam, and controlling these inputs, specifically by eliminating peak hours, seem promising approaches for IT assistance.

Requirement III: The system should integrate inputs from energy meters, and enable the machine operator to perform a fine-granular analysis of the current energy inputs and develop plans for improving energy efficiency, e.g., by eliminating peak hours. Additionally, planning information coming from e.g., an ERP system is required, meaning that the system has to provide a generic input interface.

Based on those requirements, we derived the system design presented in the next section.

6 System Design

The system is built upon the paradigm of service-oriented architectures (SOA) (Weerawarana et al. 2005). It uses a set of core services, utilizing open common Internet standards such as WSDL (W3C 2001) and BPEL (Business process execution language for web services 2003).

The system design is based on a layered architecture, separating frontend and backend services, enabling flexible adaptation of the SOA composition to the requirements of specific SME use cases, e.g., integration of specific enterprise protocols in the backend. The backend services, which are from a technical point of view the more complex part of the system, have already been implemented and tested on a wide range of test data, demonstrating the basic feasibility of our system design. The system’s frontend services are still under design, as those require inputs from the actual people going to use them to reach their full potential. The preparation of pilot studies is ongoing.

6.1 Backend Services

The system’s “intelligence” is realized on the server side by several backend services. In order to measure the real energy consumption of a production system and its components, an Energy Consumption Monitoring Service interfaces energy meters and stores the data in the backend. In order to use the stored data, further data processing is performed, such as data preparation (e.g., smoothing of measured data). For some parts of the offered decision support, forecast of future energy consumption and energy prices is needed to provide a base on which decisions, like component replacement, can take place. The service provides forecasting functionality based on past data by implementing a SARIMA model (Olsson and Soder 2008).

The Data Management Service is responsible for providing read/write access to the data and meta-data of the system. We use ontologies to describe the information from various platform-internal and external sources. This approach has already proven to be purposeful, especially in heterogeneous environments (Bullinger 2006; Bügel and Laufs 2008). The Information required by the system is structured in specialized data models (Fig. 5). For the realization of the backend models, we decided to use the web ontology language (OWL) (McGuinness et al. 2004) which is built upon the less expressive W3C standards RDF (Ankolekar et al. 2008) and RDFS (Celino et al. 2009).

Fig. 5
figure 5

System design overview

Full size image

6.2 Front End Services

When providing system functionality aimed at increasing energy efficiency to the stakeholders involved in the production process, an easy-to-use user interface is crucial. For a seamless integration into given processes, a mobile solution is desirable because it also allows to use the system directly in the production environment. Mobile systems also typically have less power consumption than stationary PCs. In order to provide an affordable solution that can amortize in SME, it is necessary to rely on products from the mass market in a low price range.

While smaller mobile devices like smartphones offer better mobility, the small dimension of the screen is of course an issue regarding the visualization of vast amounts of data e.g., for visualizations of the production system’s configuration (Fig. 6).

Fig. 6
figure 6

User interface

Full size image

In order to provide a mobile solution that can fulfil the requirements regarding the user interface but is also mobile, we decided to use tablet PCs for the implementation of the front end. Affordable tablets are available on platforms like Android and Windows. There are also versions available that have been designed for use in harsh industrial manufacturing environments.

6.3 User Interface Concept

The user interface concept divides the provided functionality into four main sections. The sections are cross-linked wherever required. Parts of the sections are disabled based on the stakeholder’s role defined in the role model.

Overview Section: This section of the user interface is designed for the support of users by providing a fast/clear overview which aggregates all relevant information on a high level. This way, detected inefficient states of the production system can be easily located and further, more detailed information or assistance can be accessed.

Assistance Tools Section: The user interface provides a specific assistance tool for each supported use case. For example, the assistance tool for exchanging components provides a specific recommendation for the exchange of a component, the system has recognized to be inefficient. While the assistance tool already provides the available information derived from measurements and forecasts, the user is able to modify all relevant parameters before performing a calculation of the time of amortisation.

Data Visualisation Section: For the more detailed information of some stakeholders, the user interface also provides additional data visualisations which include both aggregated visualisations as well as raw data visualisations e.g., diagrams containing the measured energy consumption or energy prices. For raising the awareness of energy consumed by the production system, also live data can be visualized.

Notification Section: In addition to the on-demand information provided in the other three sections, the notification section resembles a push-service that notifies the relevant stakeholders when new information or new recommendations are available. Notifications can be deactivated as needed based on the notification type. An example of a pushed notification is a “turn-off recommendation” for a specific component or system. In addition to the in-line notifications in the user interface, notifications can also be transported via e-mail or SMS to reduce the need of hardware e.g., for workers.

7 Related Work

There has been research in strategic, large-scale models, analyzing energy in- and outputs of whole value chains, and used in strategic planning of future development scenarios (Schleich 2009; Hu and Bidanda 2009) for assessing the sustainability of complex, large enterprises or entire industry sectors. The EU road mapping project IMS2020 provides a roadmap for future production research, identifying sustainable and energy-efficient production as key areas. Bunse and Vodicka (2010) offer a review of tools and performance criteria for Energy efficient production based on the project’s results.

Information systems for supporting energy efficiency within enterprises also have been investigated. Roos and Hearn (2004) present a decision support system integrating enterprise planning systems with control systems and sensors to enable control of energy inputs in the ferroalloy industry. As they note themselves, such an integrated decision support system would need a high level of commitment from management, which is unlikely in most other industry sectors due to the less relevant role of energy costs (Kentor 2003) (Requirements I & II).

Similarly, Guo and Zhang (2009) propose an agent-based system for “intelligent manufacturing”. This system would require a coordinated effort between the manufacturers, all its suppliers, and all suppliers of manufacturing equipment used. Therefore, it also does not meet our requirements. It is also notable in how it goes beyond energy efficiency, and aims at controlling entire value chains fully automatic. It seems unlikely such a system would be very attractive for the SME targeted by our approach.

Vijayaraghavan and Dornfeld (2010) present an approach for energy monitoring of machines in production, which supports and integrates analysis on various levels, from value chains to machine sub-components. Our approach is quite similar, but instead of presenting on the details of the energy monitoring model, we focus on the overall information system used to apply the model.

Bengtsson et al. (2011) propose a system based on a broad life cycle model of energy and material consumption as well as waste outputs, which are fed into a simulation component for forecasting of future. The system presented is applied to assess environmental impacts over the whole product life cycle, using Excel macros evaluated by domain experts. Our system is conceptually similar, but we present an integrated information system aimed at SME, which is less suitable for overall analysis, but offers an integrated user experience, requires less expertise, and focuses on realizing immediate benefits for SME.

8 Conclusions

The chapter describes scenario parameters as starting point to realize a value-based integration of mobile and stationary devices for energy-efficient dimensioning and energy-optimized operation within the framework of order-related manufacturing. Selected strategies for energy efficiency decisions on the shop-floor-level were taken into account. Role definitions were presented to address the key stakeholders within this context. Scenarios and role definitions set the stage for further analysis and evaluation of an assistance system to realize energy efficiency potential on shop-floor level.

A service-oriented architecture is used to allow portability of the system across different manufacturing environments. On the server side, a bundle of key services provides generic functionality like data management, sensor data fusion and state data analysis. In the system’s frontend, stakeholder interaction such as notification and system calibration is realized using a front-end library that can support several platforms, many of which are especially energy-efficient.

Applying this approach, we expect to give viable support to assist decision makers facing the trade-off between potentially long life spans of existing equipment and energy-efficiency goals, as well as to realize energy efficiency potential on shop-floor level.