A real fluid does not slip at the surface of a solid boundary. Most experimental results of a Newtonian fluid satisfy with this condition. If a real fluid can slip freely over the surface of a solid boundary, how can we deduce the slip velocity. The purpose of this study is to experimentally clarify fluid slip velocity of Newtonian fluids at the duct wall. Velocity profiles of tap water and 20 wt% glycerin solution flowing in a square duct with a highly water repellent wall were measured by means of a conical hot film anemometer. Consequently, the velocity profile with the slip is measured, and the laminar drag reduction phenomena and the friction factor formula for a square duct with fluid slip at the wall have been obtained.

View full abstract

This paper presents the results of a two-dimensional stability analysis on supersonic flow relative to a cascade, using actuator disk methods. The results show the possibility of existence of two types of destabilizing mode in the supersonic cascade flow. One mode corresponds to conventional rotating stall in subsonic cascade flows. This mode can be unstable both in subsonic and supersonic regions, and its propagation velocity ratio(absolute propagation velocity / rotor speed)is smaller than unity. The other mode is a newly found mode unique to supersonic cascade flows. This mode can become unstable only in the supersonic region, and its propagation velocity ratio is for the most cases larger than unity. Both modes become unstable due to the positive slope of the characteristic curves caused by flow compressibility as in the case of, for example, a shock wave.

View full abstract

A computational fluid dynamics model, based on χ-ε turbulence model for the closure of steady turbulent flow, was used for the evaluation of the swirling flow field. Also a vortexing fluidized bed combustor(VFBC)cold model testing facility was designed, fabricated, and tested. Computations and experimental data permitted the investigation of the effects of various operating parameters on the gas flow profiles and vortex numbers. The results show that a stable suspension layer in the freeboard was produced by the swirling gas motion. The vortex number is found to increase with the secondary air velocity, increase with the diameter of the imaginary circle, and decrease with the ratio of volumetric flow rates of the primary air to the secondary air. In this paper, a correlation for evaluating the vortex number was proposed which was shown to be one of the most useful parameters in the design of a VFBC.

View full abstract

This study presents the measured rotating stall signal patterns in a vaneless diffuser in a centrifugal compressor with a radial bladed impeller. Unsteady flow and totating stall in the vaneless diffuser were investigated by measuring unsteady velocity fluctuations at several different diffuser radius ratios and axial distances using a hot wire anemometer. The flow characteristics in terms of the radial and tangential velocity compenents and the flow angle distribution in the vaneless diffuser during rotating stall were investigated using phase-locked averaging techniques. The results clearly indentified abrupt rotating stall at several different impeller rotational speeds. According to the experimental results, two different mechanisms exist for the development of the reverse flow zone in a vaneless diffuser. One is dominated by the extension of the reentering flow from the diffuser exit, and the other is dominated by the growth of the local flow separation zone on the hub and shroud side. The fluctuation in the flow direction increases a the diffuser radius ratio increases and is dominated by the strength of the reverse flow. At the onset of rotating stall, the radial velocity for one period of the rotating stall slowly increased to a maximum value and then decreased quickly to a minimum value with an intermediate peak. However, at lower flow rates, this intermediate peak did not occur.

View full abstract

Using a highly water-repellent coating wall, new type of drag reduction for a rotating disk in Newtonian fluid have been clarified experimentally. Measurements were carried out to measure the moment of rotating disks with highly water-repellent-wall and with smooth aluminum wall. The clearance between the disk and the housing wall was varied at 5, 10, and 20 mm thick. Tested fluids were tap water and aqueous solution of glycerin at concentration of 30 wt% ahd 40 wt%. The experimental data of the moment of the rotating disk with highly water-repellent wall in housing decreased comparing with the data of smooth plane disk. It was shown that the maximum drag reduction ratio was about 45% in aqueous solution of glycerin at concentration of 40 wt%, and that the viscous properties had a slight effect upon the drag reduction.

View full abstract

In order to predict various phenomena related to flow and heat transfer in space, it is necessary to carry out experiments under microgravity. The present work is concerned with the characteristics of gas-lipuid two-phase annular flow under microgravity condition utilizing a drop tower. The experiments were carried out in a vertical transparent acrylic tube of 10.5 mm I.D.and 200 mm length, using a mixture of gaseous nitrogen(GN₂)and water as the working fluid, in an annular flow region. The mean void fraction, pressure drop and liquid film thickness under microgravity were obtained and compared with the results of a ground test. It was found that the roughness of the liquid film surface was less and the mean liquid film thickness became greater under microgravity than under normal gravity.

View full abstract

Detailed flow structure of bubbly flows observed in a bubble plume is investigated numerically using the Euler-Lagrange model where the bubbly flow is treated as a continuum and each bubble is tracked in the flow. The governing equations are formulated with emphasis on the translational motion of the bubble in non-uniform unsteady liquid flow. Improvement of the present numerical method is confirmed by comparison with the Eulerian method, and the numerical results are validated by the comparison with the two kinds of experimental results, i.e., global flow pattern visualized by direct-lighting method, and local bubble motion measured by particle imaging velocimetry. In this paper, two-dimensional bubble plume which is confined by two parallel plates is analyzed and the parametric dependency, such as influence of the bubble size distribution and void fraction, on the whole two-phase behavior is clarified and experimental validation is conducted.

View full abstract

The torsional flutter boundary of a pitching hydrofoil was experimentally investigated in the subcavitation region. Unsteady lift and moment were measured using load cells installed in a torsional vibration apparatus. The type and the degree of separation on the hydrofoil exert much influence on the unsteady fluid force and the flutter boundary in the subcavitation region. In the absence of separation, no unstable range of the flutter exists in the noncavitation or the subcavitation region. In the noncavitation region, the flutter boundary extends to the higher reduced frequency side as the angle of attack increases. In the subcavitation region, these flutter boundaries remain until the cavity length reaches the mid-chord. However, the flutter boundary shifts towards the higher reduced frequency side, keeping the width of the unstable range almost constant as the cavity length extends from the mid-chord to the full chord.

View full abstract

The transportation mechanisms of fluids and / or solid films proposed in our previous papers are achieved by the coupled waves which propagate along a thin elastic membrane covering a magnetic fluid layer in a shallow and long rectangular vessel. In order to control the propulsion system and to improve its efficiency, it is necessary to adjust the frequency and wavelength of waves by considering the dispersion relation of the coupled system consisting of a magnetic fluid layer and an elastic membrane. In this study, transient coupled waves generated by applying external impulsive magnetic forces on the concentrated small region of the magnetic fluid layer are investigated experimentally and partly analytically. Various dispersive waves are observed depending on the geometrical and physical parameters such as density of the magnetic fluid, depth of the fluid layer and kind and thickness of the membrane.

View full abstract

An experimental study has verified the dynamic equation of state derived from a thermodynamic model representing simultaneous heat and mass transfer from the lipuid supply front to the solid-liquid interface in a liquid saturated porous medium. Measurement of pore liquid pressure at the solid-liquid interface fixed on the top surface of an Ohya-tuff specimen(the porous medium)enabled us to experimentally verify the relationship between temperature, pore liquid pressure and solid pressure at the interface, and liquid supply pressure. This experimental method was established by re-verifying the so-called Clausius-Clapeyron equation(GCCE), the static equation for a porous medium without liquid flow. This study also clarified the relationship between the dynamic equation of state and the GCCE on the pore liquid pressure-liquid supply pressure plane.

View full abstract

A dynamic equation of state at the solid-liquid interface for a lipuid-saturated "solid-liquid-porous media" system has been derived from a thermodynamic study modeling simultaneous heat and mass transfer through purous media from the liquid supply front to the solid-liquid interface. This is the first step in building a mathematical model which macroscopically describes a freezing expansion phenomenon with mass transfer in saturated porous media. It has been revealed that the derived dynamic equation of state generalizes the so-called generalized Clausius-Clapeyron equation, i.e.the static equation of state for porous media without mass transfer, and that mechanical energy produced at the solid-liquid interface is equal to mechanical energy dissipation due to the mass transfer.

View full abstract

An upward plane jet impinging on the free surface of a shallow rectangular tank oscillates without any external periodic force. The movement of the impinging point leaves additional fluid mass on the surface behind the point, which does not balance the momentum supplied by the jet. The imbalance generates propagating waves, and a surface level gap appears there. The level gap is flattened not by the waves but by the vertical motion of water columns. The imbalance causes lateral displacement of jet, which in turn causes the imbalance, forming a positive feedback loop. The above model explains well why the frequency corresponds to that of water column oscillation in a partitioned tank with the same water depth, and the oscillation region has a wide range above a certain velocity limit determined by the water depth.

View full abstract

In relation to reduction and passive control of noises emitted from ducts or openings of machineries, the acoustic field radiated from a flanged cylindrical duct having a coaxial cavity at the exit of the duct was investigated for various depths and radii of the cavity. Structures of the acoustic field generated by a sound source located at a distance from the exit were analyzed numerically using the finite difference method. Also, experiments were conducted to obtain the distribution of acoustic pressure levels and acoustic intensities around the exit using a three-dimensional intensity microphone system. It is shown that the cavity attached to the duct affects the acoustic field markedly and an appropriate configuration of the cavity reduces sound pressure level significantly. It is also shown that the acoustic field is well simulated by the present numerical calculation method.

View full abstract

A train moving at high speed in an open area induces the formation of pressure distribution around itself. This pressure distribution moves with the speeding train, bringing about environmental problems. Furthermore, pressure variation increases in magnitude as train speed increases. It is therefore necessary to predict the increase in magnitude and the characteristic of pressure variation in order to successfully carry out high speed railway projects. Using the boundary element method, we investigated pressure variation brought forth by train passage. Numerical results are compared with measured results from a field test, and qualitative agreement between them is determined. We are in the process of seeking countermeasures to reduce pressure variation during train passage.

View full abstract

For aiding the design and improving the performance of various Stirling engines and coolers, a Stirling engine thermodynamic and mechanical analysis, SETMA, has been developed and examined. A simple SETMA, whose working space is divided into five control volumes, has been developed to enable easy, accurate prediction of the performance of Stirling machines. In this paper, the SETMA and fundamental equations are given in detail. Several examinations of the applicability of SETMA are discussed, comparing the calculated and experimental results of two types of actual engines. The calculation accuracy can be markedly improved by optimizing the friction-loss factors and heat transfer coefficients and carefully scanning of the dimensions of the working space. For calculations of the thermodynamic properties used by SETMA, it was clarified that the SRK equation of state is more reliable than the ideal-gas equation when analyzing higher-pressure-charged Stirling machines.

View full abstract

We have proposed a numerical calculation method for predicting the solidification caaracteristics of paraffin oil droplets(tetradecane, CH₃(CH₂)₁₂, melting point 5.8°C, diameter from 4 mm to 6 mm)ascending in a cold water solution layer(a mixture of ethyleneglycol(CH₂OH·CH₂OH)and water, temperature from -2°C to -10°C)due to buoyancy. The oil droplets started to ascend with several initial velocities and initial temperatures(10°C-60°C)in the water solution layer. The oil droplets were decelerated by hydraulic resistance of the water solution flow. The water solution flowed vertically downward at a low velocity of 0.1 m / s. The temperature of the oil droplet decreased due to direct-contact heat transfer with the cold water solution, and the oil droplet started to solidify when its surface temperature reached its freezing point. The solidified fraction of the oil droplet was calculated by taking into account a combination of the flow behavior and solidification characteristics of the oil droplet. As a result, the motion of the oil droplet, the completion height and time taken for solidification were determined. The parameters used were the diameter, initial velocity and initial temperature of the oil droplet, and temperature and velocity of the water solution.

View full abstract

A numerical simulation of a premixed flame stabilized by a bluff body was performed using the LES(Large Eddy Simulation)turbulent model and a turbulent combustion model which was developed through an evaluation of a local quenching effect due to flame stretching. The latter model also includes the temperature effect of unburned gas and the pressure effect in a combustor. Due to the strong organized motion created by the bluff body, fluctuations of the premixed flame could be seen. Furthermore, the distributions of the time-averaged velocity, temperature and intensity of the velocity fluctuation were compared with experimental data and found to be in good agreement.

View full abstract

In a previous paper[Phys.Fluids 9(1997)], we proposed an improved renormalization group(RNG)theory based on conditional iterative-averaging for turbulent flow with mean shear to derive an eddy-viscosity type turbulence model. In the present paper, we have applied this iterative-averaging RNG theory to the governing equation for a thermal field, and derived a thermal eddy diffusivity representation relevant to the turbulent heat flux. Furthermore, we have also obtained the RNG-based equation for the turbulent Prandtl number Pr_t, with a hogh Reynolds-number limit as Pr_t=0.79. The present RNG-based model is applied to the calculations of wall turbulent heat transfer. Agreement with the experiment and the direct simulation data is generally very satisfactory.

View full abstract

In our previous work, we found that the preferential diffusion in a turbulent flame played an important role in its turbulent combustion sharacteristics, and estimated the local burning velocity in premixed turbulent combustion experimentally, taking account of the preferential diffusion effect. In this study, a model, which takes the preferential diffusion effect into consideration, is proposed to predict the premixed turbulent burning velocity, using the local burning velocity as a reference instead of the original laminar burning velocity. The model can be explained as follows. The turbulence affects the turbulent burning velocity by increasing the flame surface area and stretching the flame. Consequently, the turbulent burning velocity and quenching limit are determined by the balance of both effects. The predicted velocities are compared with the measured turbulent burning velocities where the fuel, equivalence ratio and the laminar burning velocity were varied extensively. As a result, quantitative accuracy of this simple model is confirmed.

View full abstract

The flow field of a supersonic flow chemical oxygen-iodine laser is simulated by solving the three-dimensional full Navier-Stokes equations, and the dependence of the mixing / reacting zone structure and the resulting gain region on the penetration depth of I₂ jet into the primary flow of singlet delta oxygen O₂(¹Δ)is investigated. The effects of wall catalysis are discussed by introducing the surface catalytic efficiency into the wall boundary condition. It is shown that an optimum condition for the secondary I₂ jet momentum exists, and that the jet that causes a high gain penetrates the primary flow up to an intermediate depth and does not collide with the counter one. It is also shown that the molar fractions of I(²P_{1 / 2})and O₂(¹Δ)are reduced markedly on the fully catalytic wall. The deactivated oxygen molecules are engulfed by the vortices generated behind the I₂ jet, leading to the presence of a large amount of unconsumed I₂, the reduced formation of I(²P_{1 / 2})and a large negative gain region in the center of the vortices even far downstream of the nozzle blades. The present study demonstrates the importance of the choice of wall material.

View full abstract

We propose a new coal-fired combined cycle power generation system featuring dust removal of high-temperature gas and wet desulfurization at the gas turbine outlet with the aim of improving the efficiency of the conventional IGCC system and reducing its construction cost. To study its feasibility, we conducted a fundamental study on the effects of alkali metal vapors in coal-derived gas on gas turbine blades and also studied remedial technology. The results of hot corrosion tests suggest that if the alkali metal vapor concentration in combustion gas is approximately 1 ppm or less current anticorrosion coating technology provides sufficient protection in practical use. The results of alkali removal tests show that 70% or higher removal rate can be achieved, which was considered a requirement for implementing the proposed system.

View full abstract

This research is intended to study relationship between cooling design of molten carbonate fuel cell(MCFC)stack and stack / system performance quantitatively. Firstly energy balance and temperature distribution are studied in 10 kW internal reforming(IR)MCFC stack. Based on the analysis 57% of waste heat is removed by endothermic reforming reaction in the IR-MCFC. The rest of the waste heat is removed by cathode gas. Cooling by internal reforming reaction(IR-cooling)reduces flow rate of recycled cathode gas necessary for stack colling, which improves stack voltage and system performance. As for the temperature distribution, IR-colling establishes more flat distribution, which increases average stack temperature and improves stack voltage. The present IR-MCFC has not fully taken advantage of the above profits of the IR-cooling design. By increasing capacity of IR-cooling, further improvement of stack / system performance can be achieved.

View full abstract

This paper describes analysis of transitional change in pressure after a quench of field windings in a superconducting generator. We constructed a calculation model by which it is possible to predict transitional changes in pressure after a quench in the helium vessel of a superconducting generator, and performed numerical calculation. To verify the calculation model, the calculated results were compared with tested results derived from the Model Rotor, and we could confirm the accuracy of the calculation model. Applying the calculation model to the 70-MW class superconducting generator, we found that the maximum pressure in the helium vessel of the 70-MW class superconducting generator reached 1.7 MPa at the quench of the maximum field current.

View full abstract

In this study, we investigated the value of the loss coefficient, ζ₂₃, at the exit of the rotor ventilation ducts when air coolant flows from the rotor ventilation ducts to the rotor-stator air gap by measuring the distribution of static pressure on the surface of the rotor and the stator and on the wall of the cooling duct of air cooled generators, by using a scale water model. Protuberant type rotor wedge was proposed and they performed much better than the flat type of rotor wedge for ventilation cooling. From the experimental results, for the protuberant type rotor wedge, the fluid can flow from the exit of the rotor cooling duct into the air gap with a small pressure loss, since the fluid near the rotor surface is separated from the rotor surface by the triangular protuberance and the static pressure above the exit of the cooling duct is less than that in the rotor cooling duct. The experimental results of the pressure loss and the results of numerical analysis which will be reported in another paper⁽¹⁾, were in good agreement.

View full abstract

In recent years, the black liquor concentration in pulp recovery boilers has been markedly elevated to its high concentration of about 80% form its conventional level of about 70%, accompanied with improved performance of the black liquor evaporator in response to the requirement for an increased handling capacity of black liquor. This trend raised the combustion temperature in char bed, causing a problem of elevated NO_x concentration level. In our present research using a field boiler, the authors verified that the NO_x level can be decreased by designing improvements of the combustion process utilizing changes of the air distribution into the pulp recovery boiler. As opposed to the current recovery boiler where primary, secondary and tertiary air are used to complete combustion, our research was conducted on the application to the recovery boiler of the in-furnace NO_x removal system where the NO_x reduction process is designed to be enhanced in the char bed zone and in its upper zone by decreasing primary, secondary and tertiary air injected into such zones and by further injecting the fourth air into the upper zone of tertiary air to complete combustion. Results of the field boiler test showed that the NO_x emission level can be decreased by about 30% compared with the conventional level by injecting the fourth air at a ratio of 8%, demonstrating that the in-furnace NO_x removal system can also be effectively applied to the recovery boiler.

View full abstract

The high efficiency and low pollutant thermal recycling technologies for combustible refuse, for example, municipal waste, have been developed by using the circulating fluidized bed boiler. MHI's circulating fluidized bed boiler is equipped with the external heat exchanger and able to control heat adsorption rate in the combustor. Therefore, the various kinds of refuse can be used and the load swing range is very wide. As there is little corrosive gas such as HCl in the external heat exchanger, a superheater can be installed and high temperature and high pressure steam can be generated. And NO_x emission is low, DeSO_x rate by limestone is high and dioxin emission is also low, because the temperature in the combustor is homogeneously 850°C and the residence time of combustion gas is several seconds. MHI has already constructed a coal and paper sludge co-fired boiler, which has been operating very favorable. The combustion tests of municipal waste, RDF(Refuse Derived Fuel), and waste plastics etc.have been also executed by the test circulating fluidized bed furnace and good results have been obtained as expected before. In addition to this combustion technology, high temperature corrosion resistant material, pollutant gas and fly ash disposition technologies are also necessary.

View full abstract

A simulated fluidized-bed heat exchanger with tube banks was visualized using a neutron rediography system. Observed flow pattern indicated an importance of the tube arrangement. Heat transfer experiment, conducted simultaneously, indicated a close relationship between the heat transfer and flow pattern. The staggered arrangement with relatively small pitch has a potential of high heat transfer performance.

View full abstract

This paper describes the resonant response of the compound periodic structure, which includes both grouped and one-ringed blade structures in a turbine wheel to form a circumferential periodicity with a packet as a unit. The resonant condition is derived by calculating the energy suppled from engine order excitation to a natural mode composed of plural nodal diameter modes. Then, the resonant stresses are calculated by equating the supplied energy and the energy dissipated by damping. As an example the resonant response and stresses are determined for a simple blade model assuming a fixed mode firstly, and then a traveling mode around the circumference.

View full abstract

In Japan, all large capacity fossil-fuel power plants are supercritical units and a steam condition of 24.1MPag, 538 / 566°C has been adopted. Through extensive development work, the design and material technologies for steam turbines with a 593°C steam temperature have been established, and the steam condition of 24.1MPag, 538 / 593°C was applied to the 700 MW steam turbine of Hekinan No.3 Unit, Chubu Electric Power Co., Inc.for the first time in Japan. This unit has been operating successfully since it started commercial operation in April, 1993. Mitsubishi Heavy Industries, Ltd.(MHI)designed 1 000 MW class turbines with more advanced steam comditions of 600°C class main steam and reheat steam temperatures for the first time in Japan, applying the latest technologies developed for 600°C class application.

View full abstract

Corrosion and Corrosion fatigue properties of 3.5 Ni-Cr-Mo-V steam turbine low-pressure rotor steel were investigated. Immersion tests were performed on 3.5Ni-Cr-Mo-V steel by dipping the specimen into deionzed water and 10 ppm NaCl solution. Observation of SEM and X-ray cartography showed that MnS inclusions were partly dissolved and corrosion pits were initiated from the edge of the dissolved MnS. Long-term corrosion fatigue tests were done in aqueous environment up to 5×10⁸ stress cycles. The S-N curve was consisted of two different curves and there appeared substantial reduction in fatigue strength over 5×10⁷ stress cycles. The fatigue crack initiation and growth from pits can be considered as the main factor of reduction of fatigue strength. Assuming the corrosion pit as a semi-elliptic surface crack, the critical stress intensity factor range ΔΚ_th for corrosion fatigue crack initiation was calculated. The average value of ΔΚ_th was around 1.50 MPam^{1 / 2}, about one half of the value obtained from conventional macroscopic fatigue crack growth rate tests.

View full abstract

Recently, the demand for expanded natural graphite is increasing as a material for the gland packing of valvesin various industrial fields in place of asbestos. This is in response to the needs for longer service life, higher reliability and lower maintenance, as well as restrictions against the use of asbestors. The expanded natural graphite gland packing has the advantages of far less retightening works and a longer service life due to its superior features of sealing and heat resistance. However, there are disadvantages in that the mechanical strength is lower and frictional resistance in movement of the valve stem is higher. To overcome these drawbacks, we have developed a gland packing using graphite pellets of 3-6 mm in diameter made by coating natural graphite with inorganic binder on ceramic core balls of 1-2 mm in diameter based on a new concept.

View full abstract

Steam pressure loss in a tube bundle significantly influences condenser performance. Based on this fact, a new tube arrangement will be developed. The analytical study shows the advantage of this improvement : up to 20% higher heat transfer performance than from our former design⁽¹⁾. The method applied to estimate performance is a steam flow analysis using the finite volume method. To confirm the analytical results, hydraulic analogue tests with water are carried out. This new type condenser will be operated in power stations in the near future.

View full abstract

In order to effectively utilize fly ash as reproduction resources, it is necessary to qualify physical and chemical properties of fly ash and to collect an evaluated quality of ash by selection and storage. So a fly ash qualify control system capable of measuring and analyzing the component and the particle size distribution of ash in onstream was developed and a demonstration plant is now installed on site. This quality control system consists of an automatic sampling system, a residual carbon content analyzer, an elementary composition analyzer, a quality evaluation equipment and the like, and is capable of operating automatically. This paper introduces its quality evaluation technique in this system and the result of its effectively demonstration test as well as the state of operating machines.

View full abstract

Both the properties of slag and ash, and the processes of deposition and growth must be clarified in order to solve slagging trouble^{(1), (2)}. This report presents the effective thermal conductivity measuring results of ash deposit and the estimating results of ash deposit internal temperature. The experiment was carried out by accumulating molten ash onto the heat transfer measuring probe disk. The main results are as follows. 1.Thickness of sintered layer decreases as the probe disk surface temperature increases. 2.The ash deposit average porosity and the ash deposit effective thermal conductivity are correlated with the distance from probe disk surface to measuring point. 3.Temperature at the interface between sintered layer and molten layer is estimated as IDT ; initial deforming temperature of ash.

View full abstract

A method of determining the deformation amount by examining the distortion of grains in transmission electron microscope(TEM)called IGD(Intra-granular Distortion)analysis was developed. In this paper, we quantitatively evaluated creep and tensile strain by the shift vector of Kikuchi pattern. Good correlations between average distortion in grains and macro deformation amount were obtained for two heats of 2.25Cr-1Mo steel after tensile and creep deformation. Based on the results, the creep damage of a 2.25Cr-1Mo heat exchanger tube, used for 220 000 hours of service in a fossil power plant, was assessed. Creep strain of the heat exchangeer tube was estimated to be 2% by IGD analysis, which corresponded to life consumption of 70%, reasonable for the material used for 220 000 hours in service.

View full abstract