Xin Q. Diesel Engine System Design

Подождите немного. Документ загружается.

56 Diesel engine system design

1. At a given engine speed and power, assuming the air–fuel ratio and EGR

rate are known requirements, nd turbine area and EGR valve opening.

2. For a transient emissions cycle with prescribed engine speed and load

varying as a function of time, assuming the transient history of the

air–fuel ratio and EGR rate are known, nd the transient history of the

turbine area and EGR valve opening as a function of time.

Design for target is the fundamental design technique in diesel engine system

design and a stepping stone toward the more advanced nondeterministic

designs. The shortcoming of design for target is that without the information

of the probability distributions of the attribute parameters or reliable previous

experience, it is difcult to select an appropriate target value in order to

achieve a design that is neither over-designed nor under-designed. Design

for variability needs to be used to address this issue.

Design for variability

Design for variability addresses the effects of both control factors and noise

factors in order to control both the mean value and the variation range of the

attribute parameters by changing the control factors. Design for variability

includes two subsets: robust design and sensitive design, one for noise input

factors and the other for control input factors. The goal of design for variability

is to achieve a reliable design that is insensitive to noise factors but sensitive

Nominal

design

target

Nominal

calibration

target

Calibration

limit

Design limit

(failure limit)

New lab engine,

the worst cylinder

Piece-to-piece noise,

internal noise (e.g.,

in-vehicle), external

noise (e.g., weather)

Piece-to-piece noise,

change-over-time noise

(wear), customer usage

noise, internal noise (e.g.,

sensor), external noise

(e.g., weather)

Probability

Functional or attribute

value (e.g., peak

cylinder pressure)

3-sigma tolerance

of design limit

1.18 Engine system design constraints or limits.

Diesel-Xin-01.indd 56 5/5/11 11:41:29 AM



57The analytical design process and diesel engine system design

to control factors. The reliable design means that the probabilistic distribution

of the system response has a reasonable mean value and deviation range so

that a predetermined percentage of population satises the requirements of

performance, durability, or packaging without failures. The response can be

any parameter of performance, durability, or packaging.

Robustness means that the system/component response is insensitive to,

or not adversely affected by, the variation of the input noise factors within

a range of circumstances, even though the sources of variability have not

been eliminated. Robustness can be measured by the signal-to-noise (S/N)

ratio. It is the effect of noise factors that robust design wants to control, via

changes in the control factors. Robust design is the process to achieve the

dened robustness. Sensitivity analysis, mean value design (for setting the

nominal target), and tolerance design (for setting the specication range)

are three important tasks in robust design.

Robust design should not be confused with sensitive design. The engine

needs to be a sensitive system which responds quickly to the control input

factors, for example, to achieve good transient performance. Note that the

system needs to be sensitive to the control factors rather than the noise factors.

When noise factors present, it is desirable for the engine system to be insensitive

to the noises in order to ensure a stable operation under uncertainties.

There are numerous examples of robust design in engine applications. For

example, the fuel system of the diesel engine must provide a consistent supply

of liquid fuel regardless of external or internal noise factors. Inconsistent

fuel delivery may cause engine stumbles when cruising, accelerating and

decelerating, rough or rolling engine idle, and other drivability issues. Higher

temperatures in the engine may lead to fuel vaporization which may cause

insufcient fuel delivery to the injectors. A robust fuel delivery design needs to

isolate or minimize the impact of these unfavorable external temperatures.

Statistical probability distributions of random parameters are required in

design for variability. Several types of distribution can be used to model the

input factors, such as normal, uniform and Beta distributions (Table A.2).

The combination of the probability distributions of different input factors

generates a probability distribution of the output response. Appropriate mean

and standard deviations of the control factors can be searched or optimized

simultaneously in order to achieve the desirable distribution of the response.

The design solution produced by such a design-for-variability approach is

usually a more cost-effective and more robust design than that obtained by

using the deterministic design approach.

As shown in Fig. 1.19, when multiple design constraints exist, as is

usually the case with engine design, the deterministic approach cannot

completely handle design-for-limit if the limit value of only one constraint

is used because that single point does not serve as the limiting case for all

the design constraints. For example, if a design-for-limit case requires an

Diesel-Xin-01.indd 57 5/5/11 11:41:29 AM



58 Diesel engine system design

analysis at peak cylinder pressure of 200 bar, this case would give a lower

exhaust manifold temperature than a case of 180 bar cylinder pressure due

to higher air–fuel ratio. Therefore, this case cannot be used to assess the

durability of the exhaust manifold. The probabilistic approach of design for

variability does not have this limitation because it produces the statistical

distributions of all the design constraints simultaneously so that all the

design constraints can be easily identied, as illustrated in Fig. 1.19. A

multi-objective optimization can then be conducted to try to seek a solution

that satises all the constraints. Monte Carlo simulation is an effective tool

used in design for variability and this topic is detailed in Chapter 3.

Some research work related to design for variability and robust design

was reported in the literature in the areas of emissions, cooling, and vehicle

fuel economy. Yan et al. (1993) and Dave and Hampson (2003) used the

DoE method and Monte Carlo simulation to analyze the robust design of

emissions and engine BSFC. Rahman and Sun (2003) analyzed engine

cooling system design to control coolant temperatures. Catania et al. (2007)

presented a robust optimization for vehicle fuel economy analysis.

Design for reliability

Design for reliability is design for variability extended into the time-in-service

domain. One example is turbocharger matching to account for the variation

Probability

Peak cylinder pressure

Coolant heat rejection Turbocharger speed Compressor outlet air

temperature

Exhaust manifold gas

temperature

Exhaust manifold pressure

1.19 Coordination between different engine system design targets

and constraints.

Diesel-Xin-01.indd 58 5/5/11 11:41:30 AM



59The analytical design process and diesel engine system design

in exhaust restriction caused by soot loading changes in the DPF. Exhaust

restriction (the pressure drop across the aftertreatment system) signicantly

affects turbocharger matching and engine system design. When the exhaust

restriction becomes higher, the turbine expansion ratio and the engine air

ow rate usually reduce. This results in a decrease in peak cylinder pressure

but an increase in exhaust manifold gas temperature. The nominal design

point of turbocharger matching is dependent on exhaust restriction. Unlike

the mufer-only exhaust system that has a deterministic back pressure at

the rated power condition, modern diesel engines equipped with a DPF have

a randomly uctuating back pressure over time, increasing during the soot

accumulation phase or decreasing after DPF regeneration. Matching the

turbocharger for a clean DPF or a fully loaded DPF yields very different

results. If the probabilistic distribution of the exhaust restriction during the

vehicle lifetime can be found, the turbocharger can be matched better to

balance the system efciency and durability.

Like design for target, design for variability/reliability is often carried

out with optimization techniques. In reliability-based design optimization

(RBDO), in order to reduce computing time, it is useful to rst conduct a

deterministic optimization to pre-screen, and then apply the probabilistic

variations of the uncertainties to the pre-screened sub-optimal solutions for

further optimization. The RBDO usually consists of the following steps:

1. Deterministic design-of-experiments (DoEs) variable pre-screening to

identify key factors.

2. Deterministic DoE response surface t to build emulator models.

3. Non-time-dependent nondeterministic variability analysis with key

factors.

4. Time-dependent nondeterministic reliability optimization with key

factors.

5. Conrmation runs.

These optimization topics are elaborated in Chapter 3.

1.4 The concept of cost engineering in diesel

engine system design

1.4.1 Design for proﬁt and design for value

Cost is one of the four attributes of diesel engine system design. The cost of a

part is related to the design in terms of the material required and the processes

available to build and maintain it. The central theme of cost engineering is

design for prot. The corporate revenue is directly related to product market

demand, which is affected by product attributes and market price. Revenue

is basically price multiplied by demand quantity. The design of the attributes

Diesel-Xin-01.indd 59 5/5/11 11:41:30 AM



60 Diesel engine system design

is accompanied by cost. Prot is the difference between revenue and cost.

The precise denitions of these economics terms are provided by Ostwald

and McLaren (2004). Design for prot is a process with target costing where

the effects of the design are evaluated for demand, revenue, cost, and prot,

and the prot is maximized in engineering design and decision making.

The essence of design for prot is to ensure that the product meets a target

market price to allow product competition. Another important concept in

cost engineering is value. Value is dened as a function-to-cost ratio, or cost

effectiveness. Maximizing value should be also considered in design.

1.4.2 The need for engine system cost analysis

Many engine development programs failed due to ineffective control of product

cost and price. In today’s highly competitive market, customer requirements on

engine product functionality are increasingly stringent. Better fuel economy,

higher power, faster acceleration, lower noise, and longer reliability life are a

few examples. On the other hand, the increasingly booming supplier industry

is trying to provide engine manufacturers with a wide range of components

or modularized subsystems with advanced or luxury features. Two-stage

variable-geometry turbocharger, electric boosting, camless valvetrain and

hybrid powertrain are several examples. An engine product can be developed

by throwing in many advanced features simultaneously to achieve superior

functional performance, but it may not be nancially viable at all. There

is a pressing need for system cost integration and control during engine

development in order to design the least expensive product while meeting

customer requirements.

Although cost reduction opportunities exist in almost all engine components,

there are three main aspects that drive a high product cost, namely emissions

compliance, fuel economy improvement, and advanced features.

The components with major cost increase for low-emissions engines

include the following (in order from high to low):

∑ aftertreatment (DPF, LNT or lean NO

trap, HC dosing, additional

sensors about pressure, temperature and NO

, precious metal loading)

∑ turbocharger (two-stage turbocharger or VGT)

∑ EGR system (larger EGR cooler size, more coolers, more EGR valves,

intake throttle valve)

∑ fuel injector (resulting from the requirement of higher injection

pressure)

∑ cylinder head (resulting from increased cylinder pressure)

∑ piston and piston rings (resulting from increased cylinder pressure and

change in piston cooling)

∑ crankcase (resulting from increased cylinder pressure).

Diesel-Xin-01.indd 60 5/5/11 11:41:30 AM



61The analytical design process and diesel engine system design

The following technologies may incur high product cost for fuel economy

improvement:

∑ hybrid electric or hydraulic powertrain

∑ waste heat recovery with Rankine, Brayton or Sterling cycles

∑ turbocompounding

∑ variable valve actuation

∑ variable compression ratio

∑ variable swirl

∑ cylinder deactivation

∑ continuously variable transmission

∑ exible cooling.

Advanced engine features associated with high product cost may include

the following examples:

∑ high power output (affecting structural strength or engine size)

∑ demanding acceleration characteristics and transient response (related

to hardware design and software controls)

∑ powerful engine brake or driveline retarder (for heavy truck applica-

tions)

∑ in-cylinder pressure real-time monitoring and closed-loop combustion

controls

∑ NVH and noise reduction features (e.g., engine encapsulation or material

selection for low noise).

Each of the above items has a heavy weight on overall engine product

cost. If each of them is not planned and controlled carefully, the entire engine

will be at a risk of high cost.

Many engine development programs failed due to the lacking of cost

control or cost design. In diesel engine system design there is a need to plan

the integrated cost roadmap for various advanced technologies for their entire

life cycles and to coordinate/balance the cost structure of the production

design between different subsystems or components. During this cost design

process, a proper balance among the four attributes (performance, durability,

packaging, and cost) needs to be achieved. Normally, an engineer in the areas

of performance, durability, or packaging is not sufciently knowledgeable

or qualied to conduct the system cost design. The cost design requires an

in-depth and up-to-date knowledge of the competitive benchmarking product

costs for all the engine components and advanced technologies, ranging from

the overall powertrain cost of OEM products to the cost of global suppliers’

component products. The cost design also requires a broad knowledge on

the relationship between cost and the other three attributes. Moreover, the

trend of modularization and integration of subsystem designs in systems

engineering environment requires the cost engineer to examine and propose

Diesel-Xin-01.indd 61 5/5/11 11:41:30 AM



62 Diesel engine system design

the most cost-effective approaches on how to conduct the modularization

and integration. In addition, the cost engineer also needs to identify the party

in charge of the modularization and integration. For example, a supplier

may prefer to provide the engine manufacturer with a wholly integrated

subsystem of waste heat recovery (consisting of heat exchangers, turbines

and electronic controls) and charge a high price for the integrated package.

Although it may look appealing to a performance or durability engineer due

to the ease or convenience, this may not be a nancially viable option for

the engine manufacturer. The manufacturer may prefer to modularize the

waste heat recovery subsystem in a different and more cost-effective way to

different global suppliers and conduct the integration work by itself without

the need for paying a high price to others. A system performance or system

durability engineer may be capable of proposing a system design plan to

meet the functional or reliability requirements, but he or she is usually not

eligible or available to optimize the cost structure of the product. All these

cost-based activities have to be planned, conducted, and optimized by a highly

qualied cost engineer. It is obvious that such a job function of system cost

design is an independent full-time engineering function rather than a small

supplemental function attached to the design.

1.4.3 The process of design target costing

The cost design activity of a cost engineer in engine development is a part of

the overall nancial analysis for the new product development. The nancial

analysis is characterized by four main phases as explained by Menne and

Rechs (2002), namely (1) cost target calculation, (2) piece cost calculation,

(3) piece cost verication, and (4) monitoring of running costs. Product cost

design occurs from program introduction to Job 1 (start of production). The

process of design for prot starts at the top with market price and prot goals.

At the beginning of the program, from a nancial perspective, the target

product cost is set as the maximum engine cost which allows the market

price of an engine, negotiated with a customer, to achieve a certain level of

return on net asset. This nancial cost target is the fully loaded cost including

purchased materials and parts, product tooling, manufacturing labor, warranty

cost, etc. Eventually, the cost target is stripped down to the engine system

cost of total product materials and parts. The target is then translated by

the engine system design team to a technology roadmap through integration

and optimization of subsystem design costs while trying to meet customer

functional requirements. There is a negotiation and iteration involved during

the cost target setting process in order to close the gaps. Sometimes trade-

offs need to be made between functionality and cost. The initial established

target cost is then distributed among different phases and work areas, and is

cascaded downward to lower subsystem and component levels. The target

Diesel-Xin-01.indd 62 5/5/11 11:41:30 AM



63The analytical design process and diesel engine system design

cost is nalized through iterations and cost reduction activities. Product cost

control is conducted by comparing actual reported costs against estimated

costs, identifying and focusing attention on those work areas that deviate

seriously from the initial estimates.

1.4.4 Objectives of engine system cost analysis

Engineering design drives cost. The cost engineer carries out an engineering

design function for the cost–technology structure of the product. The cost

engineer is neither a nancial controller in program management nor a

nancial analyst in purchasing or nance. The cost engineer neither plans/

controls engine development labor costs, nor calculates pricing and corporate

protability. Instead, the engine system cost engineer focuses on the technical

design of the cost structure of the engine, which is mainly related to materials/

parts cost and product life. The objectives of engine system cost design/

analysis usually consist of the following six types:

1. Corporate cost–benefit analysis for various cutting-edge engine

technologies (e.g., aftertreatment, advanced combustion, alternative

fuels, fuel economy improvement, or hybrid powertrain technologies).

2. Demand–revenue–cost–prot analysis for engineering design decisions

to maximize corporate protability (i.e., design for protability).

3. Cost payback analysis to optimize between initial product cost (or

standing charge) and running cost (e.g., total cost recovery over time

through reduced fuel consumption to offset the higher initial cost).

4. Optimized unit cost structure for selecting engine technologies and

balancing subsystem costs with value-based design to meet the functional

requirements and the cost target during production engine design.

5. Life cycle cost saving estimate for design improvement.

6. Competitive benchmarking cost analysis.

1.4.5 Classiﬁcation of cost and factors in cost analysis

Total product cost can be dened as the total of unit manufacturing cost,

life cycle cost, and quality loss cost. When using cost to calculate the value

associated with a functional requirement (recall that ‘value’ is dened as the

ratio between function and cost), it is important to use the total product cost.

The unit manufacturing cost includes all the manufacturing costs such as

variable cost, xed cost and tooling cost. The life cycle cost includes all the

operational costs and the costs of warranty, repair, and routine maintenance.

The quality loss cost takes into account less tangible cost incurred by the

manufacturer, the customer, and society.

The product cost can also be broken down into two categories: variable

cost and xed cost. The variable cost includes direct material, direct labor, and

Diesel-Xin-01.indd 63 5/5/11 11:41:30 AM



64 Diesel engine system design

variable overhead (e.g., direct material or parts cost, supplier transportation

cost, warranty cost, direct manufacturing labor cost, non-wage labor cost,

manufacturing tooling cost, and overhead costs). The variable cost changes

in total dollar amount with volume. The xed cost is the one which remains

constant in the total amount of dollars throughout a specic period of time

even if there are changes in volume. Examples of xed cost include capital

expenditure and xed overheads. The xed cost is ‘diluted’ or spread over

the volume. The engineering cost can be a part of overheads, or can be

handled as a separate item if the expense is exceptionally large. Figure 1.20

illustrates the relationship among costs, price, volume, and prot.

Generally, a cost engineer considers the following factors in order to

conduct cost analysis:

∑ variable cost

∑ xed cost

∑ production methods

∑ production volumes

∑ product life (annual and life cycle mileage, number of hours or years), user

usage proles (e.g., vehicle weight, driving cycle), and operational costs

(e.g., fuel cost, oil and maintenance costs, aftertreatment running cost)

Cost or revenue (dollars per year)

Volume of production (units)

Loss

Proﬁtable

Break even

Total cost

Proﬁt

Fixed costs

Direct labor cost

Direct material cost

Variable overhead

cost

Total net

sales revenue

1.20 Illustration of cost, price, volume, and proﬁt.

Total

variable

cost

Diesel-Xin-01.indd 64 5/5/11 11:41:30 AM



65The analytical design process and diesel engine system design

∑ bill of materials (BOM), or a cost bill-of-materials tree

∑ product specications and technical drawings

∑ purchasing and supplier source/cost data

∑ interest rate, exchange rate, and ination rate

∑ depreciation.

Among the above factors, the purchased materials/parts cost and product

life are the most important ones for diesel engine system cost engineers.

Moreover, many factors contain a range of geographic or temporal variation

or uncertainties (e.g., diesel fuel price, vehicle annual accumulated mileage).

System cost analysis and estimate is not an exact science. For its statistical/

probability analysis the reader is referred to Ostwald and McLaren (2004).

1.4.6 The method of engine system cost analysis

Figure 1.21 illustrates an example of the cost payback analysis with many

inuential factors included. It is assumed that a device associated with certain

engine technology imposes an initial cost for the customer. The device could

be any arbitrary one, for example, a waste heat recovery device such as an

organic Rankine cycle system, or a diesel aftertreatment device. Engine BSFC

or vehicle fuel economy can be improved by using such a device hence the

operating fuel cost can be reduced. A straightforward discounted cash ow

methodology is used to identify the time period required to recover the

cost of the device. The interest rate and net present value are considered to

include the time value of the money. The horizontal axis in the gure gives

the anticipated annual mileage accrued for a diesel vehicle. The vertical axis

Baseline

Technology

device cost

= $3000

Diesel fuel

price = $2.5/

gallon

Vehicle fuel

economy =

9 mpg

Fuel

economy

saving =

10%

0 50,000 100,000 150,000 200,000 250,000

Annual mileage (mile/year)

Payback period (years)

Baseline deﬁnition:

Technology device cost = $1000

Diesel fuel price = $1.5/gallon

Vehicle fuel economy = 6 mpg

Fuel economy saving by the ‘device’ = 5%

Annual interest rate = 5%

Note: Any operating cost of the secondary

ﬂuid associated with the ‘device’ has not

been included in the calculations in this

ﬁgure.

1.21 Cost payback analysis.

Diesel-Xin-01.indd 65 5/5/11 11:41:31 AM

