The Impact of Variable Ambient Temperatures on the Energy Efficiency and Performance of Electric Vehicles during Waste Collection

Abstract

The market for electric cars (EVs) is growing quickly, which has led to a diversity of models and significant technological advancements, particularly in the areas of energy management, charging, range, and batteries. A thorough analysis of the scientific literature was conducted to determine the operational and technical parameters of EVs’ performance and energy efficiency, as well as the factors that influence them. This article addresses the knowledge gap on the analysis of ambient temperature-related parameters’ effects on electric garbage trucks operating in particular urban traffic conditions for selective waste collection. To optimize vehicle routes, a computational model based on the Vehicle Routing Problem was used, including the Ant Colony Optimization algorithm, considering not only the load capacity of garbage trucks but also their driving range, depending on the ambient temperature. The results show that the median value of collected bulky waste for electric waste collection vans, depending on the ambient temperature, per route is 7.1 kg/km and 220 kg/h. At a temperature of −10 °C, the number of points served by EVs is 40–64% of the number of points served by conventional vehicles. Waste collection using EVs can be carried out over short distances of up to 150 km, which constitutes 95% of the optimized routes in the analyzed case study. The research contributed to the optimal and energy-efficient use of EVs in variable temperature conditions.

1. Introduction and Background

Globally, the electric car market has grown rapidly in recent years, intensifying competition as more manufacturers offer a wider range of electric vehicle models. According to the IEA Report [1], global expenditures on electric vehicles (EVs) exceeded $425 billion in 2022, which presents an increase of 50% compared to 2021. There were 26 million electric cars on the road globally in 2022, up 60% from the previous year. According to Hardman et al. [2], this large market boom is due to financial incentives, subsidies, and regulations supporting the electrification of transport, which contribute to the increased demand for EVs, of which key factors were explored by Tu and Yang [3]. Gnann et al. [4] examined important determinants of the market diffusion of plug-in electric vehicles (PEVs), while the presentation of the effects of different efforts on the growth of distinct global marketplaces came from Zhou et al. [5] and by Wang et al. [6]. The market for battery electric vehicles (BEVs) must prioritize the development of energy charging infrastructure. Therefore, many scientific studies have been devoted to issues such as demand analysis (Zhang et al. [7] and Kabir et al. [8]), allocation optimization (Gupta et al. [9] and Yi et al. [10]), and infrastructure planning (Anastasiadou et al. [11], Sun et al. [12], and Danese et al. [13]) as well as recent trends (Deb et al. [14]) and the best practices (Hall and Lutsey [15]). The installation of charging stations on roads, in cities, and in metropolitan areas is being funded by an increasing number of nations and businesses, which increases the accessibility and convenience of using EVs. The efficiency and competitiveness of EVs in comparison to conventional internal combustion engine vehicles (ICEVs) or even hybrid electric vehicles (HEVs) are improved by these advancements.

Buses and light commercial vehicles (LCVs) with a gross vehicle weight (GVW) of less than 3.5 tons are the most widely used vehicles among businesses offering services in urban areas. Recently, a number of transport businesses have also shown interest in other trucks, such as tractor units. The registrations of electric trucks and buses by region from 2015 to 2022 are shown in Figure 1.

Figure 1 displays an IEA analysis that was published in the annual report and was based on information from the EV100, country submissions, and car insurance registration data [1]. It shows that in 2022, sales of medium- and heavy-duty trucks (which accounted for 1.2% of truck sales) and electric buses (which accounted for approximately 66,000 of all bus sales) were made globally, with China holding a significant market share.

In Europe, the share of electric buses in sales in the year under review was the highest in Finland (accounting for 75% of bus sales), in Norway and the Netherlands (almost 50%), and in Denmark (over 30%). Switzerland and Sweden were also noted for having high sales shares of electric buses. In 2022, sales of electric trucks totaled about 2000 units throughout the European Union. The majority of electric trucks are box trucks, which are medium-duty chassis cab vehicles and LCVs with an enclosed cargo area in the rear that resembles a box. When combined with their stated range of more than 300 km, EVs prove to be an effective option for last-mile deliveries on interstate routes.

The recent analysis of the largest, Chinese, market for producers of zero-emission (ZE) vehicles presented by Mao et al. [16] illustrates the great diversity in the product range. Figure 2 presents zero-emission heavy-duty vehicles (ZE-HDV) sales by segment in 2021 in the Chinese market. The electrification of the heavy transport sector is clear for many types of trucks, with the highest share of sales recorded for box trucks (15%), tractor-trailers (4%), sanitation vehicles (3%), dump trucks (2%), refrigerated trucks (0.85%), and other trucks (2%).

However, as Shen and Mao [17] show, in 2022, sales of trucks (58%), including rigid trucks, tractor-trailers, and utility trucks, outweighed sales of buses (42%) in the segmentation of the ZE-HDV market share.

One of the areas of the potentially wide use of trucks in serving the inhabitants of cities and agglomerations is the collection and transport of municipal waste.

When collecting different sorts of waste in urban and residential locations, it is crucial to assess how the surrounding temperature affects the collection vehicles’ range. For the purpose of scheduling and organizing on-demand collections, waste collection providers and local authorities need access to this data. The literature that is currently available has not looked into it in the way that this paper proposes.

Air quality in urban areas is improved and noise levels and greenhouse gas emissions are decreased when EVs are used in urban garbage management. Additionally, it lowers fleet operating expenses, which can save cities and municipalities money. Planning investments related to the use of EVs for waste collection is a complex problem that involves many factors. The operational parameters of EVs for urban waste collection must be carefully considered to ensure operational efficiency and minimize interruptions. To achieve an efficient and sustainable garbage collection operation, the specific nature of the task and the features of the local environment must be taken into account when choosing the right operating parameters for EVs used in urban waste collection. One of these factors is seasonal temperature variability, which has a significant impact on the performance, durability, and safety of batteries in electric cars.

The main research questions focus on the optimization of EVs’ bulky waste collection routes for reported calls from residents depending on the changing ambient temperature affecting the number of collection sites visited and the amount of waste. The results suggest some changes in truck routes for waste collection service providers and indicate a potential decrease in the volume of waste collected at different ambient temperatures.

The impact of dropping ambient temperatures on the quantity of waste collection sites and the mass of waste collected has not been extensively studied in the literature.

The first contribution of the paper is the application of ACO for route optimization for the collection of bulky waste for the reported collection calls from the residents and then investigating the possibility of replacement of internal combustion vehicles by EVs. The author’s program, developed at our university, was used for the optimization, while the car manufacturer’s software was used to perform the simulation. Domestic garbage has been divided into categories, some of which include commodities that can be converted from trash into fuel. Plastics, paper, and bulky garbage have the highest rates of energy recovery (Nowakowski & Wala [18]). The second important contribution of the research was the analysis of the influence of ambient temperature on the number of visited collection points, related to the energy efficiency loss in waste collection range. It allowed to calculate a potential decrease in the number of waste collection points in various ambient temperatures. Ultimately, the research contribution is important for waste collection companies that investigate the possibility of replacing the fleet of ICE vehicles in the context of the efficiency of separated waste collections depending on ambient temperature. The rest of the document is organized as follows. In Section 2, the most relevant scientific literature is reviewed, with emphasis on the technical and operational parameters of energy efficiency and performance of EVs and the factors affecting them, and presented. Section 3 presents the optimization method used in this work, based on the ACO algorithm, as an example of an application using artificial intelligence algorithms to optimize routes used by modern systems for designing travel schedules for waste collection vehicles. An on-demand collection scheme and waste collection route transport in the southern part of Poland using electric delivery vehicles with special structures were also presented. Section 4 presents the results of the efficiency of electric garbage trucks in terms of waste collection depending on changes in ambient temperature. Section 5 shows a discussion of the results with emphasis on other scientific works considering other climatic zones. Ultimately, Section 6 summarizes the primary findings from the study.

2. Scientific Literature Review

2.1. Technical and Operational Parameters on Energy Efficiency and Performance of EV

The primary benefits of employing EVs for garbage pickup are their zero emissions and reduced noise levels while in motion. These are significant elements that enhance city and urban agglomeration residents’ quality of life. Creating an intricate system for the collection and transportation of municipal waste is a challenging task. When organizing a collection, it is important to consider the operational and technological specifications of EVs. The most important elements include driving range, battery capacity, charging time and infrastructure, power and torque, strength and durability, and overall efficiency. Table 1 lists the most significant studies and scientific publications that have been published recently in this field.

One important factor in the economy’s decarbonization is the use of EVs. Waste collection firms are interested in waste collection vehicles in metropolitan communities following the introduction of electric buses, as demonstrated by Ewert et al. [19]. According to Demartini et al. [20], Dik et al. [21], and Kurniawan et al. [22], coordination of waste management, policy, and circular economy policies implemented locally are necessary for this approach to succeed.

Table 1. Summary of the most important directions of research work dedicated to the technical and operational parameters of energy efficiency and performance of EVs.

Research Parameter and Explanation	Year	Research Work	Research Description	Key Findings
Driving range—the distance an EV can travel on a single charge.	2021	Tran et al. [23]	Range extenders in BEV.	Review of various EV range-extending methods, including zinc-air batteries, internal combustion engines, fuel cells, micro gas turbines, and free-piston linear generators, with descriptions and operational principles.
	2020	Xie et al. [24]	Data from experiments based on driving range analysis and vehicle energy flow.	Driving range and energy consumption are primarily determined by the average speed of the vehicle, its running time, and the frequency distribution of the braking process.
	2021	Miri et al. [25]	Range estimation based on power-based EV energy consumption model.	Creation of a precise computer-based model to calculate the energy consumption of an EV for a specific driving cycle.
	2020	Pevec et al. [26]	Driver’s range anxiety problem.	The range anxiety phenomenon, which is the fear that an electric vehicle (EV) won’t have the driving range to get to its intended destination due to its small battery size, frequently has a detrimental impact on driving behavior.
Battery capacity—the amount of energy an EV can store.	2020	Dixon and Bell [27]	Effects on distribution networks of battery capacity, charger power, and availability for charging at various locations.	The parameters pertaining to battery capacity, charger power, and availability for charging at various locations have an impact on the ultimate charging need.
	2022	Liu et al. [28]	Advanced batteries and emerging battery technologies.	Revision and evaluation, as well as opportunities and challenges of batteries and their management technologies, are revealed.
	2021	Zhang et al. [29]	Adaptive battery capacity estimation method.	Proposition of adaptive battery capacity estimation method based on incremental capacity analysis and data-driven techniques with experimental tests.
	2021	Thingvad et al. [30]	Battery lifespan and degradation.	Development of an extensive method for measuring the battery capacity of EVs series produced via the DC charge port.
Charging time—duration required to replenish the EV’s battery to its full capacity.	2020	Kostopoulos et al. [31]	Energy losses that occur when the device is charging.	According to experimental research, the vehicle’s energy usage during charging is more than what the driver sees on the EV’s dashboard, and losses are nearly twice as high.
	2020	Chen et al. [32]	Route selection equilibrium and charging wait time equilibrium.	A bi-level mathematical model is presented to determine the best distribution of charging stations based on capacity and the balance between route selection and charging wait time.
	2020	Brenna et al. [33]	Examination of EV charging methods using converter topologies.	Based on a careful examination, the ideal charging system size is calculated, as well as potential future trends.
	2021	Karakatič [34]	Genetic algorithm-based optimization of EVs routing’s nonlinear charging times.	A revolutionary two-layer genotype genetic algorithm is proposed to reduce driving times, the number of pauses at electric charging stations, and the amount of time needed for recharging.
Power and torque—influence the EV’s acceleration, towing capacity, and overall performance	2020	Valladolid et al. [35]	Evaluation of an EV’s experimental performance using power and torque loss analysis.	Power curves for various systems include the results of the computed losses and the interpretation of the measurements; power losses are not associated with the state of charge (SOC).
	2022	Torinsson et al. [36]	Minimizing power loss in EVs by allocating wheel torque.	The first technique involves minimizing power loss through wheel torque allocation based on quadratic programming optimization, while the second method involves an offline exhaustive search.
Charging infrastructure—availability of charging infrastructure	2021	Hecht et al. [37]	EV charging station availability prediction with machine learning.	Machine learning algorithms based on historical charging station usage can be used to anticipate the occupied status of charging infrastructure.
	2021	Falchetta and Noussan [38]	Accessibility and deployment trends of the European charging network.	An analysis of the European EV charging network (with maps) using algorithms, accessibility statistics, and crowdsourced information about charging stations.
	2022	Ahmad et al. [39]	An EV charging station’s location and how it affects the distribution network.	Techniques for optimizing the distribution network’s load impact and the location of EV charging stations.
Strength and durability—related to the intensity of use, durability, and wear resistance.	2021	Vartanov [40]	High-strength steel for EVs.	The materials used to make the components for electric vehicles traditionally contain high-strength steel that was developed for cold stamping.
	2022	Gupta et al. [41]	Polymers in EVs.	Comprehensive discussion comprising newer research areas for polymers in their use for EVs.
Efficiency—longer ranges and lower energy consumption	2020	Albatayneh et al. [42]	Overall energy efficiency forEVs.	EVs supplied by natural gas power plants show the highest well-to-wheel (WTW) efficiency: 13–31%. The WTW efficiency is similar when supplied with coal-fired (13–27%) and diesel power plants (12–25%).
	2020	Weiss et al. [43]	Energy efficiency trade-offs in EVs.	The weight-related efficiency trade-offs of EVs are large and can be exploited by stimulating a shift in driving modes.

Table 1. Summary of the most important directions of research work dedicated to the technical and operational parameters of energy efficiency and performance of EVs.

Research Parameter and Explanation	Year	Research Work	Research Description	Key Findings
Driving range—the distance an EV can travel on a single charge.	2021	Tran et al. [23]	Range extenders in BEV.	Review of various EV range-extending methods, including zinc-air batteries, internal combustion engines, fuel cells, micro gas turbines, and free-piston linear generators, with descriptions and operational principles.
	2020	Xie et al. [24]	Data from experiments based on driving range analysis and vehicle energy flow.	Driving range and energy consumption are primarily determined by the average speed of the vehicle, its running time, and the frequency distribution of the braking process.
	2021	Miri et al. [25]	Range estimation based on power-based EV energy consumption model.	Creation of a precise computer-based model to calculate the energy consumption of an EV for a specific driving cycle.
	2020	Pevec et al. [26]	Driver’s range anxiety problem.	The range anxiety phenomenon, which is the fear that an electric vehicle (EV) won’t have the driving range to get to its intended destination due to its small battery size, frequently has a detrimental impact on driving behavior.
Battery capacity—the amount of energy an EV can store.	2020	Dixon and Bell [27]	Effects on distribution networks of battery capacity, charger power, and availability for charging at various locations.	The parameters pertaining to battery capacity, charger power, and availability for charging at various locations have an impact on the ultimate charging need.
	2022	Liu et al. [28]	Advanced batteries and emerging battery technologies.	Revision and evaluation, as well as opportunities and challenges of batteries and their management technologies, are revealed.
	2021	Zhang et al. [29]	Adaptive battery capacity estimation method.	Proposition of adaptive battery capacity estimation method based on incremental capacity analysis and data-driven techniques with experimental tests.
	2021	Thingvad et al. [30]	Battery lifespan and degradation.	Development of an extensive method for measuring the battery capacity of EVs series produced via the DC charge port.
Charging time—duration required to replenish the EV’s battery to its full capacity.	2020	Kostopoulos et al. [31]	Energy losses that occur when the device is charging.	According to experimental research, the vehicle’s energy usage during charging is more than what the driver sees on the EV’s dashboard, and losses are nearly twice as high.
	2020	Chen et al. [32]	Route selection equilibrium and charging wait time equilibrium.	A bi-level mathematical model is presented to determine the best distribution of charging stations based on capacity and the balance between route selection and charging wait time.
	2020	Brenna et al. [33]	Examination of EV charging methods using converter topologies.	Based on a careful examination, the ideal charging system size is calculated, as well as potential future trends.
	2021	Karakatič [34]	Genetic algorithm-based optimization of EVs routing’s nonlinear charging times.	A revolutionary two-layer genotype genetic algorithm is proposed to reduce driving times, the number of pauses at electric charging stations, and the amount of time needed for recharging.
Power and torque—influence the EV’s acceleration, towing capacity, and overall performance	2020	Valladolid et al. [35]	Evaluation of an EV’s experimental performance using power and torque loss analysis.	Power curves for various systems include the results of the computed losses and the interpretation of the measurements; power losses are not associated with the state of charge (SOC).
	2022	Torinsson et al. [36]	Minimizing power loss in EVs by allocating wheel torque.	The first technique involves minimizing power loss through wheel torque allocation based on quadratic programming optimization, while the second method involves an offline exhaustive search.
Charging infrastructure—availability of charging infrastructure	2021	Hecht et al. [37]	EV charging station availability prediction with machine learning.	Machine learning algorithms based on historical charging station usage can be used to anticipate the occupied status of charging infrastructure.
	2021	Falchetta and Noussan [38]	Accessibility and deployment trends of the European charging network.	An analysis of the European EV charging network (with maps) using algorithms, accessibility statistics, and crowdsourced information about charging stations.
	2022	Ahmad et al. [39]	An EV charging station’s location and how it affects the distribution network.	Techniques for optimizing the distribution network’s load impact and the location of EV charging stations.
Strength and durability—related to the intensity of use, durability, and wear resistance.	2021	Vartanov [40]	High-strength steel for EVs.	The materials used to make the components for electric vehicles traditionally contain high-strength steel that was developed for cold stamping.
	2022	Gupta et al. [41]	Polymers in EVs.	Comprehensive discussion comprising newer research areas for polymers in their use for EVs.
Efficiency—longer ranges and lower energy consumption	2020	Albatayneh et al. [42]	Overall energy efficiency forEVs.	EVs supplied by natural gas power plants show the highest well-to-wheel (WTW) efficiency: 13–31%. The WTW efficiency is similar when supplied with coal-fired (13–27%) and diesel power plants (12–25%).
	2020	Weiss et al. [43]	Energy efficiency trade-offs in EVs.	The weight-related efficiency trade-offs of EVs are large and can be exploited by stimulating a shift in driving modes.

One of the most crucial factors is the battery charge duration, which is significantly longer than that of combustion vehicle refueling. Important factors include the vehicle’s driving duration, the distance it can go on a fully charged battery, and the battery’s charge level.

The most important factors from the list presented in Table 1 were used as limiting conditions in this study. The limiting factor was the driving range, depending on the ambient temperature, as well as the charging time conditioning breaks in the vehicle’s operation.

2.2. Factors Affecting the Operational Parameters of Electric Vehicles

Numerous factors impact the operational characteristics of EVs, a topic of interest for numerous academics conducting studies. In previous scientific works on predicting the driving range and other operational parameters of EVs, additional variables were taken into consideration, connected with the battery degradation status (Hu et al. [44] and Farmann et al. [45]) and its health (You et al. [46], Ozkurt et al. [47], and Slattery et al. [48]) or the regenerative braking contribution to the energy efficiency (Qiu and Wang [49], Li et al. [50], and Yang et al. [51]). According to Liu et al.’s work [52] on driving heterogeneity in EVs’ energy efficiency, the aforementioned influencing elements are frequently disregarded because they are challenging to quantify and witness firsthand.

Six categories of variables—related to driver, trip, vehicle, roadway, traffic, and weather—that influence vehicle energy and emission rate were described in the comprehensive model by Ahn et al. [53].

Among the driver-related factors are behavioral variations in drivers. Hu et al.’s research findings [54] show a dynamic relationship between the driver’s declared driving style and their demonstrated behavior in a variety of traffic situations.

The distance and quantity of travels made over the course of the analysis period (such as the number of daily journeys) are examples of the travel-related factors. The goal of the study by Iwan et al. [55] was to confirm the energy efficiency of EVs when they are used to carry courier packages with frequent stops in real-world scenarios.

Numerous aspects are vehicle-related, such as its weight, size, and engine condition. Other factors include the vehicle’s air conditioning, ventilation, and catalytic converter, among other features. Berjoza and Jurgena [56] examined the distance traveled as an essential indicator affecting the exploitation cost of EVs. The exploitation cost of EVs with large capacity and weight batteries was 80–95% greater than that of EVs with low gross weight.

The roadway-related factors include the surface roughness and roadway grade, as well as the density of intersections. Travesset-Baro et al. [57] analyzed the impact of road gradients on fuel efficiency. It was determined in Liu et al.’s work [58] that taking the gradient into account increases the accuracy of estimating electricity usage from 5% to 8%.

Jonas et al. [59] looked into a number of traffic-related elements, such as the likelihood of interruptions from pedestrians or signal control, among other traffic circumstances.

The weather-related factors, which are also of interest to the research included in this study, relate to temperature, humidity, brightness, visibility wind effects, and other environmental characteristics). Studies by Fetene et al. [60], based on big data analysis, revealed how strongly the energy consumption rate (ECR) of EVs varies highly and non-linearly with driving patterns and weather conditions. The analysis of the impact of temperature on the energy efficiency of EVs was analyzed for many parts of the world, like Kuwait (Hamwi et al. [61]), the United States (Yuksel and Michalek [62]), and Alaska (Wilber et al. [63]), both for variable temperature conditions and for high temperature (Jeffers et al. [64] and Ma et al. [65]) and low-temperature conditions in winter (Hajidavalloo et al. [66], Smith et al. [67], and Aris and Shabani [68]).

This manuscript determines the impact of temperature differences on the performance of EVs in waste collection. However, there are many scientific achievements in the field of the temperature impact on BEVs. Research by Lindgren and Lund [69] indicates additional important parameters, such as the preliminary preparation and heating of vehicle cabins, which influence the optimization of fleet utilization. Detailed results of work related to cabin conditioning energy consumption of EVs range were presented by Kambly and Bradley [70], taking into account several factors of geographical and temporal differences, such as local humidity, duration and thermal soak, local solar radiation, and local ambient temperature. Similar changes in the utility of EVs are caused by the type of load that requires the use of controlled temperature or ventilation, as in the case of the climate control loads transport analysis presented by Zhang et al. [71]. Wager et al. [72] conducted additional research on the effects of employing auxiliary loads like air cooling or heating, driving EVs at highway speeds, and having a lower charge capacity due to fast-DC charging and discharge safety margins on the drivable range. In addition to the typical technical aspects related to the need for increased energy use for cooling and heating systems, aspects of the driver’s behavior and their proper training for economical driving are also important. The method outlined by Vaz et al. [73] allows the driver to plan an appropriate driving strategy by forecasting the driving range based on ideal trip parameters before the journey. The results of the work of Andreev et al. [74] indicate the need to create comprehensive system solutions enabling the creation of optimal operating conditions for batteries operating in difficult temperature conditions covering many technical and operational aspects.

Figure 3 displays the properties of the distance dependent on ambient temperature that were used for optimization in this paper.

However, it can be noticed that most of the previous analyses concern personal BEVs, not ZE-HDVs. This publication addresses a knowledge gap regarding the analysis of the effects of low temperature-related parameters on electric trucks utilized for selective waste collection in urban areas, which require specialized operational duties.

The number of potential collection places to be serviced and the collected bulky waste mass under the influence of varying temperatures were chosen from among the many criteria indicated in the literature review. The aim of the article was to indicate various parameters to systematize possible factors influencing the movement of the vehicle in any condition. All types of vehicles are affected by low temperatures in terms of movement resistance, but EVs have a particularly large drop in range.

3. Materials and Methods

Modern systems for designing travel schedules for waste collection vehicles are based on applications that use artificial intelligence algorithms to optimize routes. Based on the algorithms used, assuming the criteria of minimizing costs, fuel or energy consumption, and collection time, a collection plan is developed for a specific vehicle from the fleet. Various algorithms are used here: Arc Routing Problem—ARP (for street cleaning and collection according to the edges of the graph as a transport network) and Vehicle Routing Problem—VRP (Bramel and Simchi-Levi [75]; Ralphs et al. [76]; Ulusoy [77]) as a representation of the traveling salesman problem, i.e., waste collection from individual points represented by households. For waste collection, the limitation is the load capacity of the vehicles, both in terms of weight and volume. An extension of the routing problem is also used in the literature, Waste Collection Routing Problem—WCRP (Liang et al. [78]). Vehicle routes are optimized to reduce costs and increase energy efficiency. This includes reducing fuel or electricity consumption depending on the vehicle type.

A computational model was used to determine optimal routes, which takes into account narrow time windows for the on-demand collection of bulky waste in dispersed rural developments. On-demand waste collection requires notification of the willingness to donate waste and providing a time window for its collection. The waste collection company prepares a collection plan for a given day and informs residents about the exact collection time, presented as a narrow time window of 30 min according to the procedure shown schematically in Figure 4. Residents make the calls for waste pickup. To request the pickup of generated waste from a residence, they can contact the waste collection provider by phone or using an online form. A waste collection company’s database is optimized when calls are concluded. The drivers of each vehicle receive an optimum collection plan after the collection vehicles are assigned to them.

The general idea behind on-demand bulky waste collection is that households must phone or communicate online to arrange a service. In Poland, a garbage collecting firm provides service for a designated territory that includes one or more municipalities. Next, the collection needs to be scheduled, together with the necessary number of trucks and routes. In order to reduce expenses and improve waste collection efficiency, optimization is required. OEM simulation software is necessary to determine whether a range reduction is possible based on a number of variables. The effect of outside temperature on EV range is the main topic of our investigation.

Waste collection in municipalities usually is conducted according to a schedule. Some categories of waste can be collected on-demand by requests sent to a collection company. Then, a collection plan must be worked out with the routing of the collection vehicles. This kind of service is usually provided for the suburban areas with single-family or detached houses. Therefore, the vehicles used for the collection are vans adapted for loading bulky waste or different categories of separated waste (Figure 5). Heavy-duty vehicles are mainly used for scheduled waste collection for mixed municipal waste or from blocks of flats or residential buildings. The purpose of using vans with adapted loading space is to collect waste items or bags in any suburb including with more difficult access by larger vehicles. The battery size for the category of vehicles we used in the simulation is 52 kWh. There are two variants of quick charging capabilities with 3 or 5 h from the wall box. Quick charging mode (22 kW and 32 A) is 3 h. Rapid charging is unavailable for this model. The battery pack consists of a Li-ion nickel manganese cobalt (NMC) battery without cooling.

The transport network is mapped using nodes: transport base, waste collection points, and intersections, while road connections are edges in the graph. This is a necessary step to solve the traveling salesman problem (TSP) and the VRP (Toth and Vigo [79]).

To formulate an optimization task for the problems of routing vehicles collecting waste from households and departing from the company’s transport base, the transport network was modeled in the form of the following directed graph G:

$$G=<W,L>,$$

(1)

where:

W—a set of transport network nodes and L—a set of connections (edges) between distinguished network nodes, as access roads, with $L\subset W\text{×}W$.

The nodes where waste collection points are located are the same as the location of households. The set of numbers of households or other waste collection locations is marked as (2), where i—means the number of the i-th waste collection location and j-the number of the j-th place.

$$G=\{\mathrm{1,2},\dots i,j\dots \overline{G}\},$$

(2)

where:

$\overline{G}$—total number of households on the collection schedule.

In classical solutions, weights can be used, while the objective function presented in the equation will be transformed into a relationship that considers the criteria of minimizing time, distance, and costs. To take into account several optimization criteria, the objective function takes a value depending on the entered weights and is expressed by Formula (3).

$$F\left(X\right)={\sum }_{i=1}^k{w}_i{f}_i\left(x\right),$$

(3)

where:

k—number of objective functions, x—solution vector, and w_i—weights satisfying Equation (4).

$$w\in \left[\mathrm{0,1}\right]and{\sum }_{i=1}^k{w}_i=1,$$

(4)

Formula (3) is presented to allow for the implementation of several criteria and then a substitute criterion can be introduced, which is a weighted sum of the criteria. In the article, when determining the objective function, the collection time, collection cost, and number of points served were inspected (Nowakowski et al. [80] and [81] and Cieśla and Mrówczyńska [82]).

The article selects an optimization method using the ACO (Ant Colony Optimization) algorithm on the example of bulky waste collection carried out in Southern Poland. It has been used in several studies, achieving measurable improvements in transport efficiency (Escario et al. [83], Huang and Lin [84], and Karadimas et al. [85]). The results of using the ant algorithm for route optimization indicate an improvement in route length from 10 to 30% (Liu and He [86] and Seçkiner et al. [87]).

Using the ACO algorithm for the traveling salesman problem requires declaring the total number of ants that are in the city i. There are b_i(t) (1 = 1…n) ants between n cities with weights for the edges between vertices i and j, represented as d_ij (Dorigo [88] and Dorigo and Socha [89]). Each ant is treated as an agent that is characterized by certain properties:

• Going from city i to j stays on the edge,
• The choice of the next location on the ant agent’s route is determined by the probability, which depends on the distance between the current and the considered node and the amount of pheromone (a substance secreted by ants to mark the route) on the edge connecting both locations,
• Each town (location) can only be visited once per route.

Assuming the intensity coefficient of the pheromone concentration on the edge (i,j) at time t—τ_ij(t), each iteration of the algorithm should modify its value following Equation (5):

$$\begin{array}{c}{\tau }_{ij}\left(t+1\right)=\rho ·{\tau }_{ij}\left(t\right)+∆{\tau }_{ij}(t,t+1),\\ ∆{\tau }_{ij}\left(t,t+1\right)={\displaystyle \sum _{k=1}^m}∆{\tau }_{ij}^k(t,t+1)\end{array}$$

(5)

where $\rho$—pheromone evaporation coefficient (to limit the unlimited increase in the number of pheromones on the edge, a condition must be met: $\rho$ < 1, 1 − $\rho$—pheromone evaporation level, $∆{\tau }_{ij}^k(t,t+1)$—the size of the pheromone trail left by the kth ant on edge (i,j).

The algorithm used to carry out the optimization assumed that the pheromone was left after the entire route prepared by the ants and that its size depended on the length of the designed path.

The parameters for the ACO algorithm used a number of ants—200, a maximum number of iterations—100, and the pheromone evaporation coefficient—0.1. The calculated values of the objective functions were within the range of 0.96–0.99. The number of collection points is related to traveling time between nodes, loading time, and location of the collection company base. The optimized route was available in a short time not exceeding several seconds. The loading time was 10 min at each waste collection point.

For the routing optimization, we used the original optimization software designed at our university and investigated in Nowakowski et al. [80,81]. The application collected data about the road network and the distance matrix was created after reading Open Street Map data. The collection time at individual points was 10 min (bulky waste loading time). The optimization was carried out for typical conditions occurring in suburban waste collection, considering the speed limit zone of 50 km/h and access to individual collection points of 20–30 km/h. Traffic was not taken into account because the roads are connecting residential areas. Only access to the waste collection company base may be affected by heavy traffic.

All motion-related parameters that can be considered when calculating an EV’s range should be taken in consideration. In example, a lower temperature would result in a shorter range and increased motion resistance. Also, auxiliary devices found in EVs contribute to energy consumption, including heating the inside of a car, and can also drastically cut down the range of an electric garbage compactor truck when it uses too much energy during the loading and compacting cycles. The model for estimating the real range of a vehicle requires more data and additional measurements. It should be calculated individually for a vehicle offered by manufacturers. The case study focuses on available data from bulky waste collection in Southern Poland. The vehicles used in the collection were not equipped in auxiliary electric equipment and the results of the simulation are shown in next section of this article.

4. Results

The research was carried out for the collection of large-sized waste for single-family rural buildings (scattered) in the south of the Silesian Voivodeship in Poland. This part of the manuscript shows the results for 10 optimized routes.

Table 2. List of parameters for the implementation of collection routes for selected bulky waste.

Route Number	Route Completion Time [h]	Route Length [km]	Number of Collection Points	Collected Waste Weight [t]
1	07:42:51	254	17	1.78
2	09:16:59	285	15	1.65
3	09:15:14	222	16	1.32
4	06:30:40	179	14	1.55
5	07:10:10	195	16	1.70
6	06:40:00	181	19	1.74
7	07:40:40	176	19	1.82
8	06:30:00	171	18	1.65
9	07:40:00	183	19	1.78
10	06:35:00	186	16	1.65
Mean [-]	07:37:46	212	17	2.00
σ	√(01:32:26)	√102.69	√2.19	√0.22

presents a list of parameters for the implementation of selected bulky waste collection routes for the 10 analyzed routes.

Fifty distinct routes have been chosen in order to assess the volume of bulky waste gathered in the research area. There were between 15 and 21 collecting locations for each vehicle. The selection of these routes is included in Table 2 of the study. The statistics for the analyzed set of collections in Southern Poland are shown in Figure 6.

Each route was optimized using ACO algorithm. The distribution of the collection points depends on the boundaries of the local community. It is a dynamic schedule related to an address of the collection call, registering in a local database in a collection company, route optimization, and final assignment of a plan for the selected vehicle. A sample of the optimized route is presented on a map (Figure 7).

To optimize waste collection routes, first, a solution is generated by a greedy algorithm. The value of the objective function is always 1 and this is the reference solution. Then, the ACO algorithm starts generating results from this solution and their objective function value is determined as a quotient value describing the vehicle’s route to the value of the greedy solution. ACO terminates after 100 iterations in which it was not possible to construct a solution with an objective function value lower than the best one found.

A proposed support for the existing fleet of waste collection vehicles is EV vans for the collection of bulky waste. The vehicles are equipped with a special body to enable the loading of large items like furniture, carpets, etc. The routes shown in Table 2 were completed for the real case study for the waste collection in Southern Poland. The tests and simulations for the same planned routes were conducted using EVs by one of the manufacturers of EV vans from the EU.

The simulation and the tests for the same planned routes were conducted using EVs by one of the manufacturers of EV vans from the EU. The loading capacity of the vehicle is 13 m³, the maximum length of the cargo compartment, and GVW is 4500 kg (based on data for vehicles manufactured in 2023).

The median value of collected bulky waste for EV waste collection vehicles depending on ambient temperature per route is 7.1 kg/km and 220 kg/h. A reduced number of waste collection points depending on ambient temperature is shown in

Table 3. The number of bulky waste collection points depending on ambient temperature.

Route Number	Number of Collection Points at −10 °C	EV Range at −10 °C [km]	Number of Collection Points at 0 °C	EV Range at 0 °C [km]	Number of Collection Points at +10 °C	EV Range at +10 °C [km]	Number of Collection Points at +20 °C
1	7	88	9	113	11	138	12
2	6	90	7	105	9	135	10
3	7	88	9	113	11	138	12
4	9	104	11	127	13	150	13
5	9	90	11	110	14	140	15
6	11	92	13	108	17	142	18
7	12	100	15	125	17	142	18
8	11	97	13	115	16	141	17
9	12	106	13	115	16	141	17
10	9	90	11	110	14	140	15

. Compared to the ICE vehicle, the longest route is 285 km and the achieved waste collection time is 9 h 16 min.

The number of collection points’ number for EVs at the temperature −10 °C ranges from 40 to 64% of the number of locations for ICE vehicles. It is possible to complete almost 95% of the shorter routes, and only 70% of the longest routes, at the temperature +20 °C. The reason is the limitation of the range of EVs up to 150 km. The battery charging time is 6 h for this category of EVs. This is the main obstacle against the continuation of the collection plan after unloading in a company base. Therefore, the route ends and the additional collection must continue the next day.

EVs’ range is significantly reduced at lower ambient temperatures. The findings of this study show the effect that the lower range of a vehicle during collection requires an increased number of routes. The range of the EVs does not exceed 90 km at the temperature −10 °C and the number of collection points was decreased. At the same time, the total mass of the collected waste was also decreased. The reduction in the range of the vehicle allows for the collection of approximately 640 kg of the waste that can be used as refuse-derived fuel. Comparing the result of approximately 1700 kg of waste collected during longer routes, the efficiency is significantly lower and constitutes only 38% of the median value for the ICE vehicles.

Figure 8 shows the sensitivity analysis of the potential shortening of EVs’ range depending on ambient temperature. It includes the routes of bulky waste collections.

5. Discussion

According to earlier research (Nowakowski and Wala [90,91]), when determining whether to use EVs for waste collection in urban areas, the following variables should be taken into consideration: the vehicle’s range and the battery’s state of charge.

One of the parameters that has a direct impact on these factors is the air temperature analyzed in this manuscript, which may be highly variable in some regions. The area of Southern Poland in which the research was carried out lies in a temperate and transitional climate. We distinguish different climatic zones in the country depending on the season—winter (two zones) and summer (five zones). In summer, the analyzed area is in the second climate zone, where the calculated air temperature is 30 °C. In winter, the analyzed area is in the third zone, with the design air temperature for infrastructure facilities at −22 °C (and the annual average is 6.9 °C). The temperature range is quite large, which has a huge impact on the possibility of using specific commercial vehicles to service cities.

Using EVs for waste collection has several benefits, not the least of which is lowering exhaust emissions—a factor that is especially significant in metropolitan areas—and also lowering noise levels caused by passing cars. The short range of EVs does have some obvious drawbacks, though. This influence is large for long distances, even though it is quite minimal for short routes. Lower ambient temperatures have a far more noticeable effect on battery performance and capacity decline. As a result of decreasing range of waste collection EVs, the number of collection points is lower.

An accurate estimation of the state of the charge of the battery is crucial for battery management and vehicle performance. In low-temperature situations, the driving range may drastically decrease to roughly 60% (Al-Wreikat et al. [92]). An additional factor having an impact on SOC of the battery and the driving range is aging. In this case, capacity fades, and the resistance of a battery increases (Sun et al. [93]). It will be a subject for the continuation of our study. Some authors of scientific studies suggest even special battery thermal management systems to be adjusted for low ambient temperatures (Kirkaldy et al. [94]).

According to efficiency-related research, the impedance effect lowers a Li-ion cell’s capacity at below-freezing temperatures, increasing the cell’s internal resistance and lowering the state of charge to between 7 and 23 percent of its maximum initial SOC. (Bernagozzi et al. [95]). In some studies, focused mainly on energy consumption by EVs, it was indicated that the energy consumption of a vehicle at −10 °C was 18.7% higher when compared to that at 25 °C (Babu et al., [96]). In a different method, it was discovered that the vehicle’s energy consumption was greatly impacted by low temperatures (−7 °C). In comparison to a temperature of 25 °C, the low ambient temperature of −7 °C led to an approximate 10–18% increase in motor energy. An even higher energy consumption was indicated in a study by Lee et al. [97]. In low temperatures, the energy efficiency was limited to a 28% greater energy usage than in pleasant temperature circumstances. From summer to winter, a vehicle’s range might decrease by up to 28%. (Al-Wreikat et al. [92]).

Our study indicated a significant drop in the range of EVs and these results highlight the necessity of taking into consideration applying more vehicles for waste collection or even encapsulation of the battery using insulation material.

The research is important for waste collection companies that investigate the possibility of replacing the fleet of ICE vehicles with EVs in the context of the efficiency of separated waste collections depending on ambient temperature. The results show that it may be necessary to take into account the variability of ambient temperature in the waste collection system to plan the frequency and optimize the routes, which would take into account the decrease in the energy efficiency of vehicles. Not only planning and optimization applications, but also tools that help predict the outside temperature in the future, can be helpful.

Waste collection companies can effectively address the challenges of temperature fluctuations through the use of technology, regular equipment maintenance, and adherence to collection schedules. It is becoming necessary to use advanced waste monitoring and management systems that allow for the optimization of collection routes and response to changing conditions in real time. Practical implications are focused on understanding the decreased length of routes, and at the same time, the waste collection points visited by the collection company. It also highlights the necessity of suitable infrastructure for charging vehicles (fast chargers) or selection of vehicles for purchase with longer ranges. It is especially important in sparsely populated communities with long distances to run by waste collection EVs.

The research is limited by the challenges associated with predicting winter temperatures. This is essential for organizing waste pickup. According to our research, there is a substantial correlation between the number of collection locations visited during cold weather and the range of bulky waste collection vehicles. It helps to reduce the total mass of waste that is collected. To assess other operating factors, future research should concentrate on examining a range of automobiles from multiple manufacturers.

6. Conclusions

Electric vehicles are playing an increasingly important role in the transformation of urban transport, contributing to the creation of more sustainable, clean, and citizen-friendly environments. The development prospects of municipalities and waste management indicate further growth trends in the waste collection EVs sales market. Therefore, it is important to continue research to increase their competitiveness and efficiency, as well as to improve operational and operational parameters. The use of EVs in urban waste management contributes to reducing greenhouse gas emissions, reducing noise, and improving air quality in urban areas. Moreover, it also reduces the costs of operating a vehicle fleet, which can bring savings for cities and municipalities. It is important to consider technical support aimed at increasing the range of EVs. This includes fast chargers and the possibility of replacing batteries at the company’s base, as well as thermal insulation of batteries in vehicles.

Future research will focus on determining the operational parameters of waste collection vehicles in other temperature conditions and the impact of battery aging on the range of vehicles. A comparison of various vehicles would be a subject for a review to investigate various operational parameters for EVs manufacturers.