**3. Case Study**

In this section, we analyze in depth the case of retrofit carried out in the city of Cracow, Poland. After making an overview of the lighting system being modernized and the structure of lighting classes of relevant roads (Section 3.1), we analyze how the power usage can be decreased by selecting a given retrofit approach (Section 3.2). Next, those savings are converted into the emission reductions (Section 3.3) and extrapolated to a scale of the entire city area (Section 3.4). In the last subsection (Section 3.5), we assess how a lower CO2 emission influences an ultimate financial balance.

**Remark 2.** *It should be stressed here that the emission values presented in this section depend strongly on emission factors which vary for different countries (for details, see [21]). For the purposes of this analysis, we assumed Polish emission factors [22], which are shown in Table 4.*


**Table 4.** Emission factors for Poland.

### *3.1. The Installation's Setup*

The modernization of a lighting system covered 3768 out of 80,000 luminaires installed in Cracow. Those sodium fixtures were replaced by LED sources (see Section 2.1). Additionally, newly installed LEDs were dimmed to adjust luminous fluxes to actual needs, defined by relevant lighting classes (see Section 2.2). The charts illustrating the considered system's structure are shown in Figure 2 (the structure of lighting classes (Figure 2a) and powers of HID sources being replaced (Figure 2b)). The circuits were connected to 81 control cabinets.

**Figure 2.** (**a**) Numbers of lighting situations (in parentheses) in the retrofitted area, grouped by lighting classes. (**b**) Numbers of HID fixtures (in parentheses) being replaced in the retrofitted area, grouped by nominal power.

### *3.2. Power Usage Optimization*

The total power of the initial, sodium-based installation was *PNa* = 311.9 kW. Replacing HID fixtures by non-dimmed LEDs reduced it to *PLED* = 183.2 kW, and dimming them brought finally *PdimLED* = 157.8 kW. Average dimming per cabinet varied between 0.63 and 0.97 (where 0 denotes full dimming, i.e., a fixture being switched off, and 1 corresponds to a non-dimmed state). Frequencies of particular dimming ranges are shown in Figure 3.

**Figure 3.** The histogram representing frequencies (bar heights) of average luminaire luminous flux ratios per cabinet.

Taking into account the total annual operation time of the lighting system *T* = 4292 h [18], the above powers give annual energy usages *ENa* = 1338.7 MWh, *ELED* = 786.3 MWh and *EdimLED* = 677.3 MWh, respectively. Additionally, adding control capabilities to a lighting system reduces *EdimLED* by 27% (see Section 2.3): *<sup>E</sup>dim*,*ctrl LED* = 494.4 MWh.

### *3.3. Greenhouse Gas Emission Reductions*

Table 5 groups all data obtained in the previous subsection and presents the savings in terms of a GHG emissions reduction, as calculated according to the emission factors for Poland [22] (see Table 4). As shown, replacing HPS fixtures with LEDs gives the 41% reduction while adjusting LED dimmings brings additional 8%, thus reducing the GHG emission corresponding to sodium lamps by 49%. Introducing lighting control increases this ratio to 63%. Thus, finally, we obtain that 1 MT (metric ton) of CO2 (the same applies to other greenhouse gases) emitted when producing energy required by sodium vapor lamps is reduced to 0.37 MT of CO2 for an adjusted, controlled, LED-based lighting installation.

**Table 5.** The energy usage and GHG emissions corresponding to 3768 lighting points, for four setups: Na (sodium based–before retrofit), LED (relamping only), dimmed LEDs (after additional luminous flux tunning) and dimmed LEDs with control.


### *3.4. City-Scale Power and Emission Reductions*

Having the reductions achieved for a representative set of lighting situations (in terms of lighting classes and fixture powers) (Table 5), we can estimate the total GHG emission reduction for the entire city with 80,000 luminaires. The results are shown in Table 6.

**Table 6.** The estimated energy usage and GHG emissions corresponding to 80,000 lighting points, for four setups: Na (sodium based–before retrofit), LED (relamping only), dimmed LEDs (after additional luminous flux tunning) and dimmed LEDs with control.


Figure 4 gives a more intuitive view of the contributions brought by particular retrofit methods.

**Figure 4.** The CO2 emission reduction structure. The full circle represents emission of a sodium vapor lamp.

An interesting problem related to lighting system retrofit, in the context of GHG emission reduction, is finding the relation between a lighting class of a given area and a potential emission reduction. This relation can be obtained on the basis of statistical analysis of data gathered for a considered case. Finding this relation allows answering the question of which lighting installations should be upgraded first (e.g., in a situation of limited financial resources) to achieve the maximum environmental impact. Figure 5 presents an annual reduction of CO2 emission corresponding to the retrofit considered in this case study, broken down into contributions brought by particular lighting classes. As can be seen, the most GHG reduction was contributed by retrofitting roadways/areas of M2 and S2 classes.

**Figure 5.** The annual CO2 emission reduction (full circle) broken down into ratios contributed by particular lighting classes [13].

Figure 6 shows the above results in more tangible form, i.e., in terms of an annual CO2 emission reduction volume expressed in metric tons per 1000 luminaires.

**Figure 6.** The annual CO2 emission reduction broken down into volumes (metric tons per 1000 luminaires) contributed by particular lighting classes [13].

### *3.5. Decreased CO*2 *Emission and Financial Savings*

Let us compare now the direct financial benefits achieved thanks to the retrofit-based energy savings (i.e., reduced power consumption) and the savings derived from a reduced CO2 emission, computed on the basis of prices of European Union emission allowances (EEX EUA, or simply "EUA") traded on secondary market. The former component is calculated for the average electricity price for non-household consumers for EU-28 (second half of 2017) which equals e0.14 per kWh [23]. In turn, the EUA price is assumed to be e17.00/MT [24].

The values presented in Table 7 show that, besides the environmental impact, the CO2 reduction increases by 10% annual profits brought by a retrofit.

**Table 7.** Annual energy savings and CO2 reduction in terms of financial benefits. The average electricity price e0.14/kWh for non-household consumers for EU-28 (second half of 2017) and the EUA price e17.00/MT were assumed. Results correspond to data in Table 6.


**Remark 3.** *The following simple reasoning shows that annual financial savings (S) increased by* Δ = 10% *(CO*2*-related component) can reduce an investment payback period (T) by 9%. Indeed, if C stands for the retrofit investment costs, then T* = *C*/*S. After increasing an S value by 10%, we obtain:*

$$T' = \frac{\mathbb{C}}{\mathbb{S} + \Delta \cdot \mathbb{S}} = \frac{\mathbb{C}}{1.1 \cdot \mathbb{S}} = 91\% \\ T.$$

*It should be remarked that, since an actual EUA price is subject to stock fluctuations this ratio can change (see Table 8).*

**Table 8.** Retrofit payback growth related to CO2 reduction vs. EUA price.

