Modeling Accumulated Change

An integral records how a state changes

Many applied definite integrals begin with a state quantity and a rate. If $Q(t)$ is an amount and $r(t)=dQ/dt$ is its rate of change, then

$$Q(b)-Q(a)=\int_a^b r(t)\,dt.$$

The integral is a net change, not automatically a final amount. To recover the state at the end, include the initial condition:

$$Q(b)=Q(a)+\int_a^b r(t)\,dt.$$

This distinction prevents a common interpretation error. A flow-rate integral may give liters added to a system, but the tank’s final volume also depends on how much was present initially. An acceleration integral gives change in velocity, not velocity unless the starting velocity is included. The equation tells us exactly which quantity is being accumulated.

Units identify the meaning of the integral

Suppose $r(t)$ is a volumetric flow rate in $\mathrm{L}/\mathrm{min}$ and $t$ is measured in minutes. The product $r(t)dt$ has units of liters, and so does $\int_a^br(t)dt$. Units do not emerge after integration; they travel through the model from its first small contribution. If a rate in liters per minute is integrated with respect to seconds, a conversion factor is needed to preserve consistency.

The same logic applies broadly. Force in newtons times distance in meters gives work in joules. Density in kilograms per cubic meter times a small volume gives mass in kilograms. Probability density times an input value gives a weighted contribution whose integral has the input’s units. Dimensional analysis cannot determine every formula, but it can expose a missing thickness, a misplaced conversion, or an attempt to add incompatible quantities.

For a rate described only by a data table, numerical integration is not an inferior substitute for a symbolic antiderivative. It is the model supported by the available evidence. The method should name how intervals are approximated, state the time step, retain units in each product, and describe any error assumption. A polished decimal without a model of the unobserved behavior between samples communicates less, not more.

Sign gives direction, not a mistake

Let a one-dimensional velocity $v(t)$ be positive to the right and negative to the left. Then

$$\int_a^bv(t)\,dt$$

is displacement: rightward motion contributes positively and leftward motion contributes negatively. Equal motion in opposite directions can give zero displacement even though the object traveled a positive total distance. To calculate distance traveled, use

$$\int_a^b|v(t)|\,dt,$$

and split the integral at times where velocity changes sign if needed.

The absolute value changes the physical question. It does not “fix” a negative result in a net-change problem. A negative net flow or negative displacement may be exactly the meaningful conclusion: more volume left than entered, or the endpoint lies left of the start. Decide whether the problem asks for signed change, final state, total amount, or distance before selecting the integrand.

Build a representative contribution

The basic modeling routine is consistent across applications. First choose a variable and a thin interval, such as $dt$ or $dx$. Next express the small contribution with its units. Then identify the bounds over which the same description holds, splitting the interval if a formula or sign convention changes. Finally, integrate and interpret the result in the language of the original quantity.

Consider a medication infusion rate $r(t)$ measured in $\mathrm{mg}/\mathrm{h}$ over a four-hour interval. The delivered dose is $\int_0^4r(t)dt$ in milligrams. If an elimination rate also acts on the body, simply subtracting a separate rate integral may be appropriate only if both rates are defined for the same state quantity and time interval. A differential-equation model may be needed when elimination depends on the current concentration. The integral setup is a model statement, so its assumptions deserve words.

Checks make the model auditable

Ask whether the sign and size of the answer agree with the graph or physical description. If a positive rate dominates most of the interval, net change should be positive. If the initial state is nonnegative and the model predicts a negative final amount, investigate whether the rate law is being applied beyond its valid range. Compare a numerical answer to a rough rectangle or triangle estimate when a graph is available; order-of-magnitude checks catch misplaced units and factor-of-sixty errors.

For retrieval, a pump’s net flow rate is $r(t)=8-3t$ liters per minute for $0\le t\le4$ minutes, and the system starts with $20\,\mathrm{L}$. Set up the final-volume model, identify when the rate changes direction, and distinguish the net volume change from the total volume moved. Include units in every expression. Then explain why simply integrating $|r(t)|$ would answer a different question. The next lesson unifies convergence and approximation decisions across the course’s many representations.