Emission control status and future perspectives of diesel trucks in China.

Chinese diesel trucks are the main contributors to NOx and particulate matter (PM) vehicle emissions. An increase in diesel trucks could aggravate air pollution and damage human health. The Chinese government has recently implemented a series of emission control technologies and measures for air quality improvement. This paper summarizes recent control technologies and measures for diesel truck emissions in China and introduces the comprehensive application of control technologies and measures in Beijing-Tianjin-Hebei and surrounding regions. Remote online monitoring technology has been adopted according to the China VI standard for heavy-duty diesel trucks, and control measures such as transportation structure adjustment and heavy pollution enterprise classification control continue to support the battle action plan for pollution control. Perspectives and suggestions are provided for promoting pollution control and supervision of diesel truck emissions: adhere to the concept of overall management and control, vigorously promote the application of systematic and technological means in emission monitoring, continuously facilitate cargo transportation structure adjustment and promote new energy freight vehicles. This paper aims to accelerate the implementation of control technologies and measures throughout China. China is endeavouring to control diesel truck exhaust pollution. China is willing to cooperate with the world to protect the global ecological environment.

Optimization of municipal solid waste incineration for low-NOx emissions through numerical simulation.

With urbanization, municipal solid waste (MSW) generation is increasing. Traditional landfill methods face land shortages and environmental pollution. Waste incineration, which reduces waste and recovers resources, has become a key management method. However, nitrogen oxides (NOx) produced during incineration severely impact the environment, requiring improved control technologies. This study optimized three denitrification technologies-air staging, flue gas recirculation (FGR), and selective non-catalytic reduction (SNCR)-using numerical simulations. The research provides support for improving waste incinerator efficiency and stability while reducing NOx emissions, aiding the sustainable development of waste incineration technology. By optimizing the primary and secondary air distribution ratios, the initial NOx generation was reduced by 8.39%. When 20% of the recirculated flue gas was introduced as secondary air, NOx generation was reduced by 23.54%, and boiler efficiency increased to 83.78%. The study examined the impact of different sludge mixing ratios on the temperature and NOx emissions within the context of municipal solid waste (MSW) incineration. Initially, the study aimed to address the environmental concerns of NOx emissions during the incineration process by exploring how the introduction of sludge at various mixing ratios would affect combustion parameters. The results showed that a sludge mixing ratio between 3 and 13% optimized the combustion process with 7% being the most effective in balancing temperature control and NOx emissions. Specifically, the best value of the sludge mixing ratio refers to achieving an optimal reduction in NOx emissions while maintaining stable incinerator operation. The chemical compositions of the sludge included key elements such as carbon (C), hydrogen (H), nitrogen (N), sulfur (S), and oxygen (O), with approximate proportions of C: 31.2%, H: 4.7%, N: 2.5%, S: 0.6%, and O: 31.8%.

Evaluating methods for marine fuel sulfur content using microsensor sniffing systems on ocean-going vessels.

Establishing emission control areas (ECAs) can effectively reduce air pollution from marine emissions, making efficient monitoring of the fuel sulfur content (FSC) of ocean-going vessels (OGVs) key to ECA policy enforcement. Various FSC detection approaches, including oil sample analysis, sniffing method, and optical remote sensing, have been proposed, each with its benefits and drawbacks. Among these, the sniffing method appears promising but requires further improvement in field operation protocol and data analysis processes. This study aims to develop a comprehensive methodology, including sensor calibration, field operations, and data analysis, to enhance the performance of an Unmanned Aerial Vehicle (UAV)-based Microsensor Sniffing System (MSS) for real-time FSC monitoring. Hong Kong has a cap of 0.5 % m/m FSC for OGVs, and hence Hong Kong waters served as the "real-world" monitoring location to evaluate the MSS system through land-based and sea-based measurements. Three different FSC calculation methods were employed and verified against bunker delivery note (BDN) data through blind testing. Results confirm that the MSS is effective in field settings, though it has an underestimation tendency, demonstrating an absolute error of 0.06 % m/m, 0.11 % m/m, and 0.10 % m/m for the Crest, Slope, and Area methods, respectively, compared to BDN data. However, high errors were possible with low CO2 and SO2 peak heights, and single-peak samples compared to multi-peak samples. Over 16 successful trips, the FSC of 125 valid OGVs (Mean FSC = 0.39 % m/m) exhibited a lognormal distribution pattern, with the distribution tail approaching the 0.5 % m/m regulatory cap. This investigation highlights the potential of a UAV-based MSS for monitoring and enforcing FSC regulations within ECAs, providing a systematic protocol to guide future research direction and enforcement practices.

Pulsed processing by cold plasma, applied to industrial emission control.

A promising pollution control technology is cold plasma driven chemical processing. The plasma is a pulsed electric gas discharge inside a near atmospheric-pressure-temperature reactor. The system is energized by a continuous stream of very short high-voltage pulses. The exhaust gas to be treated flows through the reactor. The methods applied involve the development of robust cold plasma systems, industrial applications and measuring technologies. Tests of the systems were performed at many industrial sites and involved control of airborne VOC (volatile organic compound) and odor. Electrical, chemical and odor measuring data were collected with state-of-the-art methods. To explain the test data an approximate solution of global reaction kinetics of pulsed plasma chemistry was developed. It involves the Lambert function and, for convenience, a simple approximation of it. The latter shows that the amount of removal, in good approximation, is a function of a single variable. This variable is electric plasma power divided by gas flow divided by input concentration. In the results sections we show that in some cases up to 99% of volatile pollution can be removed at an acceptable energy requirement. In the final sections we look into future efficiency enhancements by implementation of (sub)nanosecond pulsed plasma and solid state high-voltage technology and by integration with catalyst technology.

Identifying Driving Factors of Atmospheric N2O5 with Machine Learning.

Dinitrogen pentoxide (N2O5) plays an essential role in tropospheric chemistry, serving as a nocturnal reservoir of reactive nitrogen and significantly promoting nitrate formations. However, identifying key environmental drivers of N2O5 formation remains challenging using traditional statistical methods, impeding effective emission control measures to mitigate NOx-induced air pollution. Here, we adopted machine learning assisted by steady-state analysis to elucidate the driving factors of N2O5 before and during the 2022 Winter Olympics (WO) in Beijing. Higher N2O5 concentrations were observed during the WO period compared to the Pre-Winter-Olympics (Pre-WO) period. The machine learning model accurately reproduced ambient N2O5 concentrations and showed that ozone (O3), nitrogen dioxide (NO2), and relative humidity (RH) were the most important driving factors of N2O5. Compared to the Pre-WO period, the variation in trace gases (i.e., NO2 and O3) along with the reduced N2O5 uptake coefficient was the main reason for higher N2O5 levels during the WO period. By predicting N2O5 under various control scenarios of NOx and calculating the nitrate formation potential from N2O5 uptake, we found that the progressive reduction of nitrogen oxides initially increases the nitrate formation potential before further decreasing it. The threshold of NOx was approximately 13 ppbv, below which NOx reduction effectively reduced the level of night-time nitrate formations. These results demonstrate the capacity of machine learning to provide insights into understanding atmospheric nitrogen chemistry and highlight the necessity of more stringent emission control of NOx to mitigate haze pollution.

Observation-Based Diagnostics of Reactive Nitrogen Recycling through HONO Heterogenous Production: Divergent Implications for Ozone Production and Emission Control.

Understanding of nitrous acid (HONO) production is crucial to photochemical studies, especially in polluted environments like eastern China. In-situ measurements of gaseous and particulate compositions were conducted at a rural coastal site during the 2018 spring Ozone Photochemistry and Export from China Experiment (OPECE). This data set was applied to investigate the recycling of reactive nitrogen through daytime heterogeneous HONO production. Although HONO levels increase during agricultural burning, analysis of the observation data does not indicate more efficient HONO production by agricultural burning aerosols than other anthropogenic aerosols. Box and 1-D modeling analyses reveal the intrinsic relationships between nitrogen dioxide (NO2), particulate nitrate (pNO3), and nitric acid (HNO3), resulting in comparable agreement between observed and simulated HONO concentrations with any one of the three heterogeneous HONO production mechanisms, photosensitized NO2 conversion on aerosols, photolysis of pNO3, and conversion from HNO3. This finding underscores the uncertainties in the mechanistic understanding and quantitative parametrizations of daytime heterogeneous HONO production pathways. Furthermore, the implications for reactive nitrogen recycling, ozone (O3) production, and O3 control strategies vary greatly depending on the HONO production mechanism. On a regional scale, the conversion of HONO from pNO3 can drastically enhance O3 production, while the conversion from NO2 can reduce O3 sensitivity to NOx changes in polluted eastern China.

Substituent effects of halogens on the excited-state intermolecular proton transfer reactions.

Fluorescent aromatic urea compounds undergo excited-state intermolecular proton transfer (ESPT) in the presence of acetate anions to produce an excited state of the tautomer (T*) from the excited state of the complex (N*), resulting in dual fluorescence. Herein, we performed spectroscopic measurements of anthracen-1-yl-3-phenylurea derivatives with substituents, -CF3, -F, or -Cl, at the p-position of the phenyl group in the presence of acetate to investigate the substituent effects on the ESPT reaction and the deactivation processes of N* and T*. Kinetic analysis showed that the reverse ESPT rate constant (k-PT) depended on the respective substituents, suggesting that each substituent may influence the reverse ESPT process differently. In particular, since the electron-withdrawing properties of -F are estimated by the - I and + Iπ effects, it is plausible that -F has a slight electron-donating property and influences the reverse process from T* to N* in the excited state. This study shows that it is possible to control emission by selecting specific substituents in the ESPT system.

Comprehensive effect of increased calcium content in coal on the selenium emission from coal-fired power plants: Combined laboratory and field experiments.

Coal combustion is the major contributor to global toxic selenium (Se) emissions. Inorganic elements in coals significantly affect Se partitioning during combustion. This work confirmed that the calcium (Ca) in ash had a stronger relationship with Se retention at 1300 °C than other major elements. Ca oxide chemically reacted with gaseous Se, and its sintering densification slightly affected Se adsorption capacities (44.45 -1840.71→35.17 -1540.15 mg/kg) at 300 - 1300 °C. Therefore, Ca in coals was identified as having potential for hindering gaseous Se emissions, and coals with increased Ca contents (2.74→5.19 wt%) were used in a 350 MW unit. The decreased Se mass distribution (3.54%→2.63%) in flue gas at air preheater inlet (320 -362 °C) confirmed the effectiveness of increased Ca content on gaseous Se emission reduction. More gaseous Se further condensed and was chemically adsorbed by fly ash when passed through an electrostatic precipitator, resulting in a significant increase in the Se content of fly ash. Additionally, the corresponding Se leaching ratio decreased from 4.88 - 35.74% to 1.87 - 26.31%, indicating enhanced stability of Se enriched in fly ash. This research confirmed the feasibility and environmental safety of sequestration of gaseous Se from flue gas to fly ash by increasing the Ca content in coals.

Evidence and causes of recent decreases in nitrogen deposition in temperate forests in Northeast China.

High reactive nitrogen (N) emissions due to anthropogenic activities in China have led to an increase in N deposition and ecosystem degradation. The Chinese government has strictly regulated reactive N emissions since 2010, however, determining whether N deposition has reduced requires long-term monitoring. Here, we report the patterns of N deposition at a rural forest site (Qingyuan) in northeastern China over the last decade. We collected 456 daily precipitation samples from 2014 to 2022 and analysed the temporal dynamics of N deposition. NH4+-N, NO3--N, and total inorganic N (TIN) deposition ranged from 10.5 ± 3.5 (mean ± SD), 6.1 ± 1.6, and 16.6 ± 4.7 kg N ha-1 year-1, respectively. Over the measurement period, TIN deposition at Qingyuan decreased by 55 %, whereas that in comparable sites in East Asia declined by 14-34 %. We used a random forest model to determine factors influencing the deposition of NH4+-N, NO3--N, and TIN during the study period. NH4+-N deposition decreased by 60 % because of decreased agricultural NH3 emissions. Furthermore, NO3--N deposition decreased by 42 %, due to reduced NOx emissions from agricultural soil and fossil fuel combustion. The steep decline in N deposition in northeastern China was attributed to reduced coal consumption, improved emission controls on automobiles, and shifts in agricultural practices. Long-term monitoring is needed to assess regional air quality and the impact of N emission control regulations.

Using a bottom-up method to assess cruise ship activity impacts on emissions during 2019-2020 in China.

COVID-19 has had a serious impact on the development of the global shipping industry, especially since the impact on the cruise tourism industry was unprecedented. This study took cruise ships sailing in China ECA, China Exclusive Economic Zone (EEZ), Yangtze River main line, and Xijiang River main line Chinese waters as an example to analyze the key changes in cruise ship emissions during the pandemic. Automatic identification system (AIS) data, vessel static data, and emission control regional data are used to conduct a comprehensive analysis of cruise ship emissions from multiple perspectives such as a port-to-regional comparison. As such, a vessel emission model (i.e., a bottom-up method) is constructed in this research for predicting China ECA and EEZ cruise ship emissions. Compared with 2019, the cruise activities sailing in China's Emission Control Area (ECA) are mainly at berth, and the emissions of cruise ships have dropped significantly, with SOx emissions reduced by 59.11%. In addition, this study also calculates the carbon emissions of China's regional cruises, supplementing China's cruise carbon pool. The research results suggest that cruise operators may improve fuel efficiency, decrease vessel speed, improve routing and scheduling, and enhance fleet management in order to further mitigate the negative effects of the cruise tourism industry on the marine environment.

Bibliometric analysis on mercury emissions from coal-fired power plants: a systematic review and future prospect.

Coal-fired power plants (CFPPs) are one of the most significant sources of mercury (Hg) emissions certified by the Minamata Convention, which has attracted much attention in recent years. In this study, we used the Web of Science and CiteSpace to analyze the knowledge structure of this field from 2000 to 2022 and then reviewed it systematically. The field of Hg emissions from coal-fired power plants has developed steadily. The research hotspots can be divided into three categories: (1) emission characterization research focused on speciation changes and emission calculations; (2) emission control research focused on control technologies; (3) environmental impact research focused on environmental pollution and health risk. In conclusion, using an oxygen-rich atmosphere for combustion and installing high-efficiency air pollution control devices (APCDs) helped to reduce the formation of Hg0. The average Hg removal rates of APCDs and modified adsorbents after ultra-low emission retrofit were distributed in the range of 82-93% and 41-100%, respectively. The risk level of Hg in combustion by-products was highest in desulfurization sludge (RAC > 10%) followed by fly ash (10% < RAC < 30%) and desulfurization gypsum (1% < RAC < 10%). Additionally, we found that the implementation of pollution and carbon reduction policies in China had reduced Hg emissions from CFPPs by 45% from 2007 to 2015, increased the efficiency of Hg removal from APCDs to a maximum of 96%, and reduced global transport and health risk of atmospheric Hg. The results conjunctively achieved by CiteSpace, and the literature review will enhance understanding of CFPP Hg emission research and provide new perspectives for future research.

Precision forecasting of spray-dry desulfurization using Gaussian noise data augmentation and k-fold cross-validation optimized neural computing.

Perceptron models have become integral tools for pattern recognition and classification problems in engineering fields. This study envisioned implementing artificial neural networks to forecast the performance of a mini-spray dryer for desulfurization activities. This work adopted k-fold cross-validation, a rigorous technique that evaluates model performance across multiple data segments. Several ANN models (21) were trained on data obtained from sulfation conditions, including sulfation temperature (120 °C-200 °C), slurry pH (4-12), stoichiometric ratio (0.5-2.5), slurry solid concentration (6%-14%) as the feed input and sulfur capture as the response. Three hundred synthetic datasets generated using the Gaussian noise data augmentation underwent a 10-fold cross-validation process before simulation on neural networks triggered by the logsig and tansig activation functions. The computation accuracy was further evaluated by altering the number of hidden cells from 2 to 10. The ANN architectures were assessed using statistical metrics such as mean square error (MSE), root mean square error (RMSE), mean absolute percentage error (MAPE), and the coefficient of determination (R2) techniques. Overall, error estimation suggests cross-validation and data augmentation are critical in efficient neural network generalization. The logsig function trained with 10 hidden cells presented closer data articulation when mapped onto actual values.

Effects of emission control areas on sulfur-oxides concentrations--Evidence from the coastal ports in China.

The setting of Sulfur limitations in Emission Control Areas (ECAs) is a crucial action of marine environmental governance at the international regulatory levels. In this study, the overall and structural impacts of the two rounds of ECA policies on SOx concentrations were quantified using synthetic control method (SCM) based on time-series data from Chinese coastal ports from 2005 to 2020. From the outcomes, the 1st round of ECA policy announced in 2015 intensified the competition between ECA and non-ECA ports and provided strong support for ECA expansion and enhanced regulation in 2019. In addition, the restrictions on the Sulfur content of marine fuels under the 1st round of ECA policy has only effectively reduced the SOx concentration in the Bohai Rim and the Yangtze River Delta region, whereas the impact on the Pearl River Delta region isn't significant. However, the 2nd round of ECA policy has only effectively impacted the Bohai Rim. In general, the effect of the 1st round of ECA policy is better than that of the 2nd round, which is mainly because the favorable effect of the further expansion of ECA policy is offset by a significant increase in vessel activity in Chinese coastal ports.

Control priority based on source-specific DALYs of PM2.5-bound heavy metals by PMF-PSCF-IsoSource model in urban and suburban Beijing.

To determine the priority control sources, an approach was proposed to evaluate the source-specific contribution to health risks from inhaling PM2.5-bound heavy metals (PBHMs). A total of 482 daily PM2.5 samples were collected from urban and suburban areas of Beijing, China, between 2018 and 2019. In addition to the PMF-PSCF model, a Pb isotopic IsoSource model was built for more reliable source apportionment. By using the comprehensive indicator of disability-adjusted life years (DALYs), carcinogenic and noncarcinogenic health risks could be compared on a unified scale. The study found that the annual average concentrations of the total PBHMs were significantly higher in suburban areas than in urban areas, with significantly higher concentrations during the heating season than during the nonheating season. Comprehensive dust accounted for the largest contribution to the concentration of PBHMs, while coal combustion contributed the most to the DALYs associated with PBHMs. These results suggest that prioritizing the control of coal combustion could effectively reduce the disease burden associated with PBHMs, leading to notable public health benefits.

Multi-objective optimisation model of a low-cost path to peaking carbon dioxide emissions and carbon neutrality in China.

A low-cost path system for achieving carbon neutrality in China was modelled using multi-objective programming by integrating industrial production, electric power, heating, transportation, and forest carbon sequestration. We aimed to minimise the total system cost, CO2 emissions, and air pollutants. The constraints included China's targets of peaking CO2 emissions before 2030; achieving carbon neutrality before 2060; ensuring industry, power, heating, and transportation supplies; promoting green energy; and implementing emission control. The model accounted for industries with high coal consumption, such as steel and chemical industries. Various power sources were considered, including coal-fired, nuclear, wind, and solar energy. Forest carbon sink and carbon capture and storage technologies were employed to achieve the emission reduction goals. The model, which was validated using available research data, offered cost-effective path schemes and exhibited high validity. Our findings emphasise the importance of structural adjustments and emission control, with electric power, heating, and transportation sectors showing higher feasibility and providing greater contributions to achieving carbon neutrality than other industries. Conversely, industrial transformation in sectors such as iron and steel, chemical, and construction materials had low feasibility and limited contribution. The modelling outcomes provide valuable insights for developing low-cost, carbon emission-targeted transportation structures in China's complex system. The results presented here demonstrate the global applicability of this method in contributing to plans aimed at meeting key carbon reduction targets.

Effects of impregnation sequence on the NH3-SCR activity and hydrothermal stability of a Ce-Nb/SnO2 catalyst.

Hydrothermal stability is crucial for the practical application of deNOx catalyst on diesel vehicles, for the selective catalytic reduction of NOx with NH3 (NH3-SCR). SnO2-based materials possess superior hydrothermal stability, which is attractive for the development of NH3-SCR catalyst. In this work, a series of Ce-Nb/SnO2 catalysts, with Ce and Nb loading on SnO2 support, were prepared by impregnation method. It was found that, the NH3-SCR activities and hydrothermal stabilities of the Ce-Nb/SnO2 catalysts significantly varied with the impregnation sequences, and the Ce-Nb(f)/SnO2 catalyst that firstly impregnated Nb and then impregnated Ce exhibited the best performance. The characterization results revealed that Ce-Nb(f)/SnO2 possessed appropriate acidity and redox capability. Furthermore, the strong synergistic effect between Nb and Sn species stabilized the structure and maintained the dispersion of acid sites. This study may provide a new understanding for the effect of impregnation sequence on activity and hydrothermal stability and a new environmental-friendly NH3-SCR catalyst with potential applications for NOx removal from diesel and hydrogen-fueled engines.

Investigation of Plant-Level Volatile Organic Compound Emissions from Chemical Industry Highlights the Importance of Differentiated Control in China.

The chemical industry is a significant source of nonmethane volatile organic compounds (NMVOCs), pivotal precursors to ambient ozone (O3), and secondary organic aerosol (SOA). Despite their importance, precise estimation of these emissions remains challenging, impeding the implementation of NMVOC controls. Here, we present the first comprehensive plant-level assessment of NMVOC emissions from the chemical industry in China, encompassing 3461 plants, 127 products, and 50 NMVOC compounds from 2010 to 2019. Our findings revealed that the chemical industry in China emitted a total of 3105 (interquartile range: 1179-8113) Gg of NMVOCs in 2019, with a few specific products accounting for the majority of the emissions. Generally, plants engaged in chemical fibers production or situated in eastern China pose a greater risk to public health due to their higher formation potentials of O3 and SOA or their proximity to residential areas or both. We demonstrated that targeting these high-risk plants for emission reduction could enhance health benefits by 7-37% per unit of emission reduction on average compared to the current situation. Consequently, this study provides essential insights for developing effective plant-specific NMVOC control strategies within China's chemical industry.

Co-drivers of Air Pollutant and CO2 Emissions from On-Road Transportation in China 2010-2020.

Co-controlling the emissions of air pollutants and CO2 from automobiles is crucial for addressing the intertwined challenges of air pollution and climate change in China. Here, we analyze the synergetic characteristics of air pollutant and CO2 emissions from China's on-road transportation and identify the co-drivers influencing these trends. Using detailed emission inventories and employing index decomposition analysis, we found that despite notable progress in pollution control, minimizing on-road CO2 emissions remains a formidable task. Over 2010-2020, the estimated sectoral emissions of VOCs, NOx, PM2.5, and CO declined by 49.9%, 25.9%, 75.2%, and 63.5%, respectively, while CO2 emissions increased by 46.1%. Light-duty passenger vehicles and heavy-duty trucks have been identified as the primary contributors to carbon-pollution co-emissions, highlighting the need for tailored policies. The driver analysis indicates that socioeconomic changes are primary drivers of emission growth, while policy controls, particularly advances in emission efficiency, can facilitate co-reductions. Regional disparities emphasize the need for policy refinement, including reducing dependency on fuel vehicles in the passenger subsector and prioritizing co-reduction strategies in high-emission provinces in the freight subsector. Overall, our study confirms the effectiveness of China's on-road control policies and provides valuable insights for future policy makers in China and other similarly positioned developing countries.

Humic-like Substances (HULIS) in Ship Engine Emissions: Molecular Composition Effected by Fuel Type, Engine Mode, and Wet Scrubber Usage.

Humic-like substances (HULIS), known for their substantial impact on the atmosphere, are identified in marine diesel engine emissions obtained from five different fuels at two engine loads simulating real world scenarios as well as the application of wet sulfur scrubbers. The HULIS chemical composition is characterized by electrospray ionization (ESI) ultrahigh resolution mass spectrometry and shown to contain partially oxidized alkylated polycyclic aromatic compounds as well as partially oxidized aliphatic compounds, both including abundant nitrogen- and sulfur-containing species, and clearly different to HULIS emitted from biomass burning. Fuel properties such as sulfur content and aromaticity as well as the fuel combustion efficiency and engine mode are reflected in the observed HULIS composition. When the marine diesel engine is operated below the optimum engine settings, e.g., during maneuvering in harbors, HULIS-C emission factors are increased (262-893 mg kg-1), and a higher number of HULIS with a shift toward lower degree of oxidation and higher aromaticity is detected. Additionally, more aromatic and aliphatic CHOS compounds in HULIS were detected, especially for high-sulfur fuel combustion. The application of wet sulfur scrubbers decreased the HULIS-C emission factors by 4-49% but also led to the formation of new HULIS compounds. Overall, our results suggest the consideration of marine diesel engines as a relevant regional source of HULIS emissions.

Safe deep reinforcement learning in diesel engine emission control.

A deep reinforcement learning application is investigated to control the emissions of a compression ignition diesel engine. The main purpose of this study is to reduce the engine-out nitrogen oxide (NOx) emissions and to minimize fuel consumption while tracking a reference engine load. First, a physics-based engine simulation model is developed in GT-Power and calibrated using experimental data. Using this model and a GT-Power/Simulink co-simulation, a deep deterministic policy gradient is developed. To reduce the risk of an unwanted output, a safety filter is added to the deep reinforcement learning. Based on the simulation results, this filter has no effect on the final trained deep reinforcement learning; however, during the training process, it is crucial to enforce constraints on the controller output. The developed safe reinforcement learning is then compared with an iterative learning controller and a deep neural network-based nonlinear model predictive controller. This comparison shows that the safe reinforcement learning is capable of accurately tracking an arbitrary reference input while the iterative learning controller is limited to a repetitive reference. The comparison between the nonlinear model predictive control and reinforcement learning indicates that for this case reinforcement learning is able to learn the optimal control output directly from the experiment without the need for a model. However, to enforce output constraint for safe learning reinforcement learning, a simple model of system is required. In this work, reinforcement learning was able to reduce NOx emissions more than the nonlinear model predictive control; however, it suffered from slightly higher error in load tracking and a higher fuel consumption.