Innovative Lighting Systems: Opportunities for Energy Savings

This paper analyses, from an energy flow perspective, the implementation of smart lighting systems in street lighting, where lights are dimmed to adapt to the flow of objects passing in a street. The research focus on the sustainability perspective of implementing a transition to smart lighting systems when compared to regular LED lighting. To account for externalities, the energy flow was addressed considering the extra electronic devices used in a smart lighting system (controllers, motion sensors, radars, and computers). To compare both traditional LED street lighting and smart lighting the paper started with a model of a 2.5-kilometre street, scaling up scenarios of the commune Ecublens, in the Swiss canton of Vaud, and then to half and all residential streets of Switzerland were examined to understand if the gains in energy savings are scalable. The research shows that, even with the additional electronic devices, the smart lighting system reduces the energy consumption of street lighting, even when considering the production of the extra components used. Financially, the extra costs of implementing smart lighting systems are offset by the savings in electricity consumption. Therefore, smart lighting systems for street lighting can be an environmentally and economically beneficial project to implement.


Introduction
In the light of the United Nations Sustainable Development goals, it is not only needed to combine efforts of public policies, corporations, and civil society but also actions from the public sphere to minimize waste and optimize processes, where public entities play not only the role of regulating and monitoring. Boons (2009) affirms that sustainable development requires system changes in both production and consumption.
The development and success of those goals depends on innovation rather than continuing the activities like usual. The Oxford Dictionary defines innovation as "A new method, idea or product", where usually when relating to tackling global or societal challenges, the expression of social innovation, which Stanford defines as: "A novel solution to a social problem that is more effective, efficient, sustainable, or just than current solutions. The value created accrues primarily to society rather than to private individuals", is used. Clean technology innovations address global problems, generating similar or better outcomes.

Smart
Considerin to underst Ecublens,   If an item is de r the energy p y to lumens em nsor, they com of light. Theref an be considere nario Figure 4 el. Figure 5 sh its intensity to    The energy input from the process 'distribution grid' to the processes of the luminaires was compared, as both systems are steady-state systems, the imports equal the exports. Therefore, it doesn't matter whether the imports or exports are compared to each other. The imports of the conventional system are 18,921.60 megajoules per year, whereas the imports of the smart lighting system are 16,183.03 megajoule per year. The comparison of these two figures or flows already indicates a difference in energy supply of approximately 2,738.57 megajoules per year for a street with two intersections. The implementation of the smart system in the second scenario of whole Ecublens goes along with a decrease of about 321,725.84 megajoules in energy supply per year. The impact of an implementation of the smart light system in these four scenarios based on the assumptions done indicates, that an increase of efficiency in energy use could be achieved.

Sensitivity Analysis on Variable 'ItemPass'
These results are based on a consistent number of passing items during the autonomous mode of the smart lighting system. Since the number of passing items has an essential impact on the energy consumption of the radar and the luminaire, Table 7 shows the results of a sensitivity analysis of the variable 'ItemPass' on the total energy consumption (TEC) in four different scenarios. An increase of passing items goes along with a higher energy consumption due to a longer duration at an intensity of 100%, which requires the full performance of the luminaire. The scenarios 1, 2, and 3 shows that the energy savings change marginally. The fourth scenario, which illustrates an increase of 100% of the variable 'ItemPass', leads to a decrease of almost 5.98% in energy savings. An increase of 100% is equal to one item passing every 2.5 minutes.
The results of the four scenarios examined strengthen the hypothesis, that a long-term study and a detailed simulation such as a Monte Carlo Simulation is necessary to develop for each residential district an appropriate model. Adapting the model of a 2.5 kilometres street with two intersections for each residential street of Ecublens or whole Switzerland is not detailed enough to meet the local circumstances and needs, which vary from location to location.

Energy Flow Along the Value Chain
The energy savings identified for the four scenarios goes along with an additional use of electronical devices. These electronical devices require a huge amount of energy during their production. From an environmental point of view, the additionally emerging energy consumption during their production must be considered, if it comes to an overall assessment of the energy use along the value chain. Zundritsch and Hewes (2017) calculated the additional energy consumption during the production.  For instance, these two scenarios including the energy consumption occurring during the production of the additionally used electronical devices shows, that the integral assessment according to the energy flow leads in case of switching from a conventional lighting to a smart lighting system still to energy savings.

Methods
To better understand the lighting consumption, the variation in sunlight during the year was analysed. Since the distribution grid only sends the electrical signal when it's dark, it is relevant to understand the hours of night-time during the year, to better assess the electricity input.
To calculate the yearly and monthly sunlight time (to calculate the night-time), firstly the latitude for the region of Ecublens 46.5296 o was identified. Afterwards the solar hour, the theta (which is a ratio of pi, the day per hour divided by the number of days in a year), the solar declination (the angle in radians which represents the solar angle towards the earth) was calculated. Based on these results it was possible to define the cosZ (which is a function of the latitude, solar declination and solar hour) and the arcossin of cosZ. In case of the arcossin of cosZ is higher than 90 o , it is night and if it is lower, sunlight is available. The data computed shows the sunlight angle per hour for each day of the year to be able to calculate the days and months of the Julian calendar.

System Model
The unique characteristic of cleantech innovations such as the smart lighting system exists in providing light on demand. Both energy flow analysis models don't contain any type of energy storage processes and are therefore not dependent on time. These circumstances imply constant flows and stocks: The first step in the definition of the system boundaries was identifying the different types of districts; residential, commercial, industrial, pedestrian, and school districts. Due to cities heterogenous characteristics the model must be broke down to a smaller scale, which represents only one type of district to be very close to reality. According to Schreder's data, smart public lighting systems are most useful when implemented in residential or industrial districts. An exact example of the desired neighbourhood is outlined in red and can be seen in the Quartier du Croset. Based on this residential street with help of GIS and Google Maps to calculate the average spacing between the streetlamps to get the number of streetlights used. Figure 8 and table 9 shows the important parameters determining the systems characteristics. consumption between 86 and 279 W. In our case, the Ampera Midi is used to meet the demands of a residential district including commercial businesses for households needs. Moreover, using the middle-class product Ampera 'Midi' allows us to meet the demands of a commercial street as well as the demands of a residential district. And consequently, enables an analysis of different scenarios of Scaling up more realistic.

Primary Data Collection
Besides the semi-structured interview, another source of primary data collection was counting passing items at a residential street at two different days between 10pm and 11pm (Table 10) to have a more realistic number of passing items during the autonomous mode, which is operating during the deep 'sleeping hours'. In our model, the number 12 was taken for a more conservative calculation. Another primary source of data was the analysis on solar hours in Ecublens (explained in chapter 2).

Secondary Data Collection
To bring the energy flow model as close as possible to the real consumption data, the consumption of the processes was calculated according to developed equations, which model the energy consumption dependent on its influences: a) Energy consumption of the smart lighting system processes: LED luminaire, radar, controller, internal server and computer Energy Consumption of the conventional lighting system: LED luminaire b) Financial: the costs of the smart lighting system divided per luminaire (the data was gathered from Philips and Silver Springs, both providers of smart lighting systems. The product costs CHF 331.40 (Maddox, 2016), respectively CHF 572.00 (Silver Springs 2013) for the one from Silver Springs. The average price was used). For all the calculations the most expensive model for the calculation was taken. Also, energy scope provided the pricing for the regular LED lighting system, for the same luminaire, with this data was possible to calculate the price difference between the regular and smart system and infer the same pricing difference for Philips. c) Data from the energy price in Vaud (Switzerland) per kWh for industries, with high energy usage. d) CO2eq: The carbon equivalent emission per kWh of energy produced in Switzerland

Calculation of Flows in the Smart Lighting System
Breaking the energy consumption down to each process of the smart lighting system, there is: a) The radar: Its consumption is not a constant value. It is rather dependent on the passing items (people, vehicles, animals, etc.). With every passing item, the radars consumption increases. Function of the process: detecting a moving item to communicate the detection to the luminaire to prepare for an increase in intensity b) The motion sensor is already included in some of the luminaires and have the same function as the radar but are not so precise. The energy consumption of the motion sensors is so low, that it can be neglected.
c) LED luminaire: The energy consumption is constant or linear depending on its adjusted mode d) Controller (LuCo): The energy consumption of the controller, which is situated on every luminaire, communicates with the controllers of the other luminaires and the radars. It is running throughout the year to send requested data for the light management. e) Segment Controller is required for every 150 luminaires to have more preciseness in communication.
f) Computer: the computer/tablet serves to monitor the system, intervene (by adjusting the settings for example), to check for maintenance problems, or to see data (e.g., energy consumption) g) Internal Server: The internal server compiles and processes the data between the controllers Starting at the starting point of the chain of cause and effect leads to firstly computing the consumption of the radars and motion sensors. Considering, that the radars energy consumption changes with passing items, its consumption was calculated according to the equation 3. On the other hand, the installed motion sensors are integrated in some of the luminaires. The motion sensors energy consumption is not dependent on the quantity of items detected, but rather a constant value and is in our case neglectable because of its very low energy consumption.
The system operates in two modes. The fixed settings mode is during the 'Rush hours' between 4pm-10pm and 6am-9am. Whereas the autonomous mode is working during the 'deep sleeping hours' between 10pm and 6pm. The value of 'ECL y ' is equal to the value of the flow 'Energy Supply' to each luminaire.
The transmission and heat losses are also known as dissipation losses. Those losses are described as an additional energy consumption, which is used by the luminaire to provide light. The amount of energy wasted is calculated as the percentage of the energy demand of each LED luminaire. Calculated according to the equation 9.
= × , % The street lighting LED losses of energy is around 30 and 35 percent (Ette et al., 2009). The arithmetical mean between 30 and 35 percent was used to calculate the emitted losses by the luminaire; 32.5%. The value of the losses is equal to the flows 'Energy Losses' leaving the process of a luminaire. By setting up the equations, the value of the flow 'Energy emitted' leaving the process of the luminaire, was calculated.
The controller used in this system is the LUCO NXP, which controls the LED driver and the ballasts. It is the main process, which is responsible for the communication between the luminaires, sensors and radars in the autonomous mode. The performance of the controller varies between 0,7W and 0,8W. The controller is working throughout the year with a constant value of energy consumption. For this model, the mean of 0,75W was used.
The equation 10 shows the calculation of the energy consumption for the controller.

ECC y = Energy consumption of controller per year [Wh]
The system is in steady-state and works even when LED lights are off (constantly receiving input to remain or change the state). Consequently, the input flows must be equal to the output flows. Therefore, the value of ECC y is equal to the flows 'Energy Supply' and 'Energy emitted'.
The segment controller has the same function as the controller described before. It contributes to more preciseness in communication between the electronical devices. This segment controller is used for every 150 luminaires. In case of having 151 luminaires two segments controller are required. The energy consumption for each segment controller is the same of the LUCO NXP: × 24 × 365 (11)

ECSC y = Energy Consumption of a segment controller per year [Wh]
The in-and output flow 'Energy Supply' and 'Energy Emitted' of the controller are equal to the value of 'ECSC y '.
As for the data storage and management between controllers there is a server, whose energy consumption is dependent on the amount of data generated by the controllers. The server's main purpose is to provide the lighting management system, which offers adjusting the settings and modes of each luminaire. Estimating the energy consumption of the server was challenging due to a broad range of servers available. In our case a server with a performance of 50W (Joel Hruska, 2012), is used. The performance of 50W is only achieved if the server is fully load. Assuming a load of approximately 30% lead to the following equation 12.
= 50 , = 0.3 × 50 × 24ℎ × 365 For usage of the lighting management system a computer in form of a tablet or PC is needed. Considering that the latter is more portable, considering a tablet for the management, with an iPad 2 from Apple. Its performance is 3,16W with display on and 0,45W in sleeping mode (Apple, 2012 All equations explained were used to calculate the values of the energy flows and can be found in detail in the annexes.

Calculation of Flows in the Conventional LED Lighting System
For comparability of both systems, for the conventional LED the same luminaire was used with the only difference, that it is operating without the additional electrical devices such as motion sensors, controllers, radars, computers, and servers.
Due to the fact, that no electronical devices are used, the only way to operate the luminaires is throughout fixed settings. The intensity is constant at 70%. Based on that the energy consumption of a luminaire was computed according 14 and 15: The flow 'Energy Supply' is equal to the value of ' ′. The transmission and heat losses were calculated according to equation 16 and are equal to the value of the flow 'Energy Losses' leaving the process 'luminaire'.

=
× 32.5% 100 (16) Setting up the equation (Annex) for each process of the luminaires leads to the values of the flows 'Emitted Energy' leaving the process 'luminaire'.

Assessment Indicators of Resource Efficiency
The two energy flow models of the conventional LED lighting system and the smart lighting system facilitate an analysis by comparing the two models with the help of assessment indicators. A commonly used assessment indicator is the transfer coefficient. The transfer coefficient describes the partitioning of the input flow of the substance energy within a process x in output flow j.
Because of the issue, that the luminaire used is the same in the conventional and the smart lighting system, the energy losses due to transmission and heat losses are also of the same magnitude. Moreover, having a steady-state system requires to assess the resource efficiency on another basis than on the transfer coefficient.
Another indicator, which facilitated to assess the efficiency in energy is to compare the energy input from the process 'distribution grid' to the processes of the luminaires and in case of the smart system also to the additional electronical devices. A direct comparison of the energy flow 'Energy Supply' between the two systems was performed.

Scaling up: Opportunity for Savings
The purpose of the case study on lighting systems is to illustrate the development of the energy flows in different scaling up scenarios. It was considered that the average number of lamps and sensors would be equal to the case study boundaries of 2.5 km.
The equation for calculating the quantities needed for each scaling up scenario is done according to the following equations: The internal server's capacity needed stays constant, hence it is independent from the number of luminaires jms.ccsenet.org Journal of Management and Sustainability Vol. 12, No. 1;2022 needed, the increased amount of processing power needed was translated in increasing servers used. For computers was used the same methodology with one more table per 1000 luminaires. Tables 11 and 12 shows the number of items needed for the conventional and smart lighting system respectively.
When considering costs, since both luminaires are the same, and ceteris paribus for all other conditions, the cost of replacement parts should be the same. The maintenance cost is lower on the interoperable network because the problem is already known through the interface, this reduces maintenance time. The employee/hour dedicated to managing the system through the interface, was the same as the reduction in maintenance costs.
For the system cost it was multiplied the most expensive luminaire (CHF 572 and CHF 400 for smart and conventional luminaires) by the number of luminaires in each of the four models, adding the electricity consumption per year in kWh, that was multiplied by the electricity cost for industrial users in Vaud of CHF 0.1496 per kWh (Romande Energie, 2015). Then it was calculated the difference between the smart and conventional lighting costs, and modelling, with the Excel solver tool for which year it would become zero (through the parameters where the years had to be bigger or equal than zero, and smaller than 23 years, all calculations and solver can be seen at the Excel spreadsheet 'Excel Model' sent as attachment), to calculate the ROI, later it was calculated on 23 years the return.

Sensiti
One of the mode deci was perfor the variabl As the v consumpti

Cons
In case of the additio into consid

Discuss
The energ considerin the energy is generall One must centres, w (SWOT) f results.   The system provides a reduced carbon footprint in both scenarios due to the reduction in energy, which also, through savings, mitigate the price difference between the luminaires, thus generating a return on investment, from the difference on savings. Another strength is the perception of the adopter on innovation, which may increase the awareness of the municipality and attract more inhabitants.
With governmental entities that might have debt, and the shift towards more energy saving products becomes more challenging, the financial return and image from innovation might attract cities, when already in the process of shifting to LED lighting, to choose the smart lighting system.
Another opportunity, which can be derived from the money savings in operating along the whole lifetime of a pole is the possibility to outsource the costs through a Public Private Partnership (PPP) or bank loans, whom might have benefits in investing in a more expensive system to profit from the energy savings. The smart lighting system might also create new sources of revenue, for example, Los Angeles partnered with the telecom provider AT&T to provide internet to clients through the poles, with the company increasing its coverage and paying for the city to use the infrastructure, the company is also creating electrical vehicles charging stations on poles (Maddox, 2016). The energy flow model might not account for all externalities, such as the energy used for producing the electronic devices used to monitor the system, or for the internet production or data centres to store long-term data collection. Another variable is the number of items passing during the autonomous mode, since energy reduction is achieved through a decrease of passing items, it is therefore necessary to analyse for each street the major characteristics of district, the circumstances, and the inhabitants needs to enable a customized analysis and implementation of smart lighting systems.
As for the electronic devices used, when assessing energy consumption, the energy savings achieved from the usage of smart lighting in 23 years outpace the energy used for the devices production, but as for materials, some resources used are scarce (e.g., lithium). From the environmental point of view the production of the electronic devices must be not only analysed from a material or an energy flow perspective, but rather also from a water usage perspective. The extraction of noble earths included in these electronical devices expend a huge amount of water. In times of water scarcity, this issue should be also studied to derive an integral positive environmental assessment of smart lighting. A lifecycle and full supply chain analysis, including the transition costs, regarding the material flow of a smart lighting transaction could be object of further studies.
Considering that smart lighting systems (both autonomous and interoperable) have digital communications implied, the systems could be hacked, impacting on the city illumination. It's necessary to increase digital security of those systems regularly to prevent attacks. Another threat is the long-term project status, where cities might not be willing to engage in long term activities and changes in government might create conflicts for the provider, even where contracts are in place.

Conclusion
This project focused on evaluating the energy consumption of LED street lighting, comparing smart and conventional systems, to assess how clean technologies innovation address the issues they propose to solve. It's shown that smart lighting systems reduces the overall energy consumption even though there are more electrical equipment installed and in use, versus traditional lighting systems.
With countries trying to reduce, through policy making, the use of energy, LED lighting for streets might be used as a new alternative, and as this research shows, the energy savings from the smart lighting systems pays the price difference from the conventional LED lighting system off through the savings in electrical consumption. Considering finances, some cases show the possibility of increasing the source of revenue, through WiFi connection on poles, or electrical vehicles charging stations, which brings even further financial results of adopting this new technology. the lights can be optimal and flows of items passing somehow controlled, like paths and parks.
For achieving a sustainable status, one must account for externalities. For the smart lighting system to be possible it's necessary to have more equipment and materials involved, and this needs to be taken into consideration. When going through the supply chain and considering the energy consumption for the production and usage phases of smart and regular LED lighting systems, the dimming from smart lighting still have an overall reduction in energy consumption, which shows that the transition to smart lighting system is overall more environmentally sustainable.

Oxford
Dictionary. Appendix A

Calculations on the Smart Lighting System
Step 1 The night hours are calculated as a mean for each of the four seasons according to the following equation:
Insert (17)  The flow 'Energy Emitted' leaving each lumianaire, which represents the emitted light of the luminaires was calculated through setting up the equations for the in-and output flows for a luminaire.
Step 8 The energy supply and consumption for each controller (LuCo) on a luminaire:  The value of 'ECC y ' equals the flows 'Energy Supply' and 'Energy Emitted' for each radar i={1,2,3,4} Step 9: The energy supply and consumption for the segment controller:  The value of 'ECSC y ' equals the flows 'Energy Supply' and 'Energy Emitted' of the process segment controller Step 10: The energy supply and consumption per year: The value of 'ECCom y ' equals the flows 'Energy Supply' and 'Energy Emitted' of the process computer Step 11: The energy supply and consumption of the internal server:

Copyrights
Copyright for this article is retained by the author, with first publication rights granted to the journal.
This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/).