The Central Idea
A scientific model connects measurable quantities. A general model may be written as
\[ \boxed{y=f(x_1,x_2,\ldots,x_n)}. \]The formula tells us how an output depends on one or more inputs. To use it correctly, we must know what the variables mean, what units they use, what domain is physically meaningful, and which quantities are treated as constants.
Variables Carry Meaning
In pure algebra, a letter can be treated abstractly. In an applied model, the letter has a role. For example, in
\[ d=rt, \]the symbols may represent distance, speed, and time. The equation becomes scientifically useful only when those interpretations and their units are clear.
Units Are Part of the Algebra
Suppose speed is measured in kilometers per hour and time in hours. Then
\[ \frac{\text{km}}{\text{h}}\cdot\text{h}=\text{km}. \]Unit cancellation confirms that the product has distance units. Unit checks often reveal setup errors before any numerical calculation is completed.
Dimensional Consistency
Quantities that are added or set equal must have compatible dimensions. An expression such as
\[ \text{length}=\text{length}+\text{length} \]may be dimensionally consistent, while
\[ \text{length}=\text{mass}+\text{time} \]is not meaningful unless additional definitions or conversion factors have been introduced.
Constants Can Have Units
In
\[ y=kx, \]the proportionality constant need not be unitless. Its units must convert the units of the input into the units of the output:
\[ \boxed{[k]=\frac{[y]}{[x]}}. \]Direct Proportionality
If \(y\) is directly proportional to \(x\), then
\[ \boxed{y=kx}. \]The ratio
\[ \frac{y}{x} \]remains constant and equals \(k\).
Example: Calibrate a Direct Model
A sensor output \(V\) is directly proportional to applied force \(F\). When \(F=40\), the output is \(V=2.0\).
\[ 2.0=k(40). \] \[ k=0.05. \]Therefore,
\[ \boxed{V=0.05F}. \]At \(F=70\),
\[ V=0.05(70)=3.5. \]Linear but Not Proportional
A model can be linear without being directly proportional:
\[ \boxed{y=mx+b}. \]Direct proportionality is the special case \(b=0\). A nonzero intercept can represent an offset, baseline, or calibration bias.
Example: Linear Calibration with an Offset
A device reads \(2\) when the true input is \(0\), and \(12\) when the input is \(5\). Assuming a linear model,
\[ m=\frac{12-2}{5-0}=2. \]The intercept is \(2\), so
\[ \boxed{y=2x+2}. \]Inverse Proportionality
If \(y\) varies inversely with \(x\),
\[ \boxed{y=\frac{k}{x}}. \]The product remains constant:
\[ \boxed{xy=k}. \]Example: Inverse Model
Suppose \(T\) varies inversely with \(r\). If \(T=12\) when \(r=5\), then
\[ 12=\frac{k}{5}, \] \[ k=60. \]Therefore,
\[ \boxed{T=\frac{60}{r}}. \]Joint and Combined Variation
Joint variation can have the form
\[ \boxed{y=kxz}, \]while combined variation may include both direct and inverse factors:
\[ \boxed{y=k\frac{xz}{w}}. \]These models appear whenever several inputs scale an output multiplicatively.
Power Laws
A broad family of scaling models has the form
\[ \boxed{y=kx^n}. \]The exponent \(n\) determines how strongly the output responds to a change in \(x\).
Scaling a Power Law
If \(x\) is multiplied by factor \(s\), then
\[ y_{\text{new}}=k(sx)^n=s^nkx^n. \]Therefore,
\[ \boxed{\frac{y_{\text{new}}}{y_{\text{old}}}=s^n}. \]Example: Square-Law Scaling
If \(y=kx^2\) and \(x\) doubles, then
\[ \frac{y_{\text{new}}}{y_{\text{old}}}=2^2=4. \]Therefore,
\[ \boxed{\text{the output quadruples}}. \]Example: Cube-Law Scaling
If \(y=kx^3\) and \(x\) triples, then
\[ \frac{y_{\text{new}}}{y_{\text{old}}}=3^3=27. \]Thus the output becomes
\[ \boxed{27\text{ times as large}}. \]Geometric Scaling
For geometrically similar objects scaled by linear factor \(s\):
\[ \boxed{L\to sL}, \] \[ \boxed{A\to s^2A}, \] \[ \boxed{V\to s^3V}. \]Example: Enlarge a Similar Object
A design is enlarged by factor \(1.5\).
Area factor:
\[ 1.5^2=2.25. \]Volume factor:
\[ 1.5^3=3.375. \]Therefore,
\[ \boxed{A_{\text{new}}=2.25A,\qquad V_{\text{new}}=3.375V}. \]Surface-Area-to-Volume Scaling
Surface area scales as \(s^2\), while volume scales as \(s^3\). Therefore,
\[ \frac{A_{\text{new}}/V_{\text{new}}}{A/V} = \frac{s^2}{s^3} = \frac1s. \]Thus the surface-area-to-volume ratio scales inversely with linear size:
\[ \boxed{\frac{A}{V}\propto\frac1L}. \]Density as a Ratio Model
Density is modeled by
\[ \boxed{\rho=\frac{m}{V}}. \]Rearranging gives
\[ \boxed{m=\rho V} \]and
\[ \boxed{V=\frac{m}{\rho}}. \]Example: Density
A sample has mass \(540\text{ g}\) and volume \(200\text{ cm}^3\).
\[ \rho=\frac{540}{200}=2.7. \]Therefore,
\[ \boxed{2.7\text{ g/cm}^3}. \]Pressure as a Ratio Model
A simplified pressure relationship is
\[ \boxed{p=\frac{F}{A}}. \]For fixed force, pressure varies inversely with area. Doubling the area halves the modeled pressure.
A Linear Response Model
Some systems are approximated over a useful range by
\[ \boxed{F=kx}. \]If \(F=250x\), with force in newtons and displacement in meters, then at \(x=0.04\),
\[ F=250(0.04)=10. \]Therefore,
\[ \boxed{F=10\text{ N}}. \]An Electrical Linear Model
A familiar algebraic relationship is
\[ \boxed{V=IR}. \]If resistance \(R\) is fixed, voltage is directly proportional to current. Solving for current gives
\[ \boxed{I=\frac{V}{R}}. \]Example: Rearrange and Evaluate
If \(V=12\) and \(R=4\), then
\[ I=\frac{12}{4}=3. \]Therefore,
\[ \boxed{I=3}. \]Inverse-Square Models
Some idealized relationships have the form
\[ \boxed{y=\frac{k}{r^2}}. \]If distance is multiplied by \(s\), the output is multiplied by
\[ \boxed{\frac1{s^2}}. \]Example: Inverse-Square Scaling
If distance triples,
\[ y_{\text{new}} = \frac{k}{(3r)^2} = \frac19\frac{k}{r^2}. \]Therefore,
\[ \boxed{y_{\text{new}}=\frac19y_{\text{old}}}. \]Empirical Models
Not every useful formula is derived from first principles. An empirical model may be chosen because it fits measured data over a useful range. Algebra can then estimate parameters, make predictions, and compare model values with observations.
Calibrating a Known Power Law
If the model form is
\[ y=kx^n \]and one nonzero point is known, then
\[ \boxed{k=\frac{y}{x^n}}. \]Example: Calibrate a Square-Law Model
Suppose \(y=kx^2\) and \(y=18\) when \(x=3\).
\[ 18=k(3)^2. \] \[ 18=9k. \] \[ k=2. \]Therefore,
\[ \boxed{y=2x^2}. \]Estimating a Power from Scale Data
If \(y=kx^n\), and multiplying \(x\) by \(s\) multiplies \(y\) by \(M\), then
\[ M=s^n. \]Therefore,
\[ \boxed{n=\frac{\ln M}{\ln s}}. \]Example: Infer the Exponent
Doubling \(x\) multiplies \(y\) by \(8\).
\[ 8=2^n. \]Therefore,
\[ \boxed{n=3}. \]Ratio Method for Power Laws
If
\[ y_1=kx_1^n \]and
\[ y_2=kx_2^n, \]dividing eliminates \(k\):
\[ \boxed{\frac{y_2}{y_1}=\left(\frac{x_2}{x_1}\right)^n}. \]Example: Ratio Method
Suppose \(y\propto x^2\), \(y_1=20\) at \(x_1=4\), and \(x_2=10\).
\[ \frac{y_2}{20} = \left(\frac{10}{4}\right)^2 =6.25. \] \[ y_2=125. \]Therefore,
\[ \boxed{125}. \]Dimensionless Ratios
Ratios of measurements with the same units are dimensionless. For example,
\[ \frac{L_2}{L_1} \]has no length unit because the units cancel. Scale factors and normalized quantities are commonly dimensionless.
Scale Models
A linear scale of
\[ 1:50 \]means one unit on the model represents \(50\) of the same units on the full object. The model-to-full-size linear multiplier is \(50\).
Example: Linear Scale
A model length is \(8\) cm at scale \(1:25\).
\[ 8(25)=200\text{ cm}=2\text{ m}. \]Therefore,
\[ \boxed{2\text{ m}}. \]Area and Volume Scale Factors
A linear scale of \(1:10\) gives an area scale of
\[ \boxed{1:100} \]and a volume scale of
\[ \boxed{1:1000}. \]Example: Scale-Model Volume
A model tank has volume \(2\) L at linear scale \(1:5\).
The volume multiplier from model to full size is
\[ 5^3=125. \]Therefore,
\[ 2(125)=250\text{ L}. \] \[ \boxed{250\text{ L}}. \]Scientific Notation
Scientific and engineering quantities are often written as
\[ \boxed{a\times10^n}, \]where
\[ 1\le|a|<10. \]Example: Multiply Scientific Notation
Multiply
\[ (3\times10^5)(4\times10^{-2}). \]Multiply coefficients and add powers of ten:
\[ 12\times10^3=1.2\times10^4. \]Therefore,
\[ \boxed{1.2\times10^4}. \]Order-of-Magnitude Checks
Estimate the scale of an answer before calculating precisely. For example,
\[ (2\times10^6)(5\times10^3) \]should be on the order of
\[ 10^{10}. \]An answer near \(10^2\) would immediately signal an exponent error.
Percent Error
If modeled or measured value \(M\) is compared with reference value \(R\), a common absolute percent-error formula is
\[ \boxed{\text{percent error}=\frac{|M-R|}{|R|}\times100\%}. \]Example: Percent Error
A model predicts \(98\) while the reference value is \(100\).
\[ \frac{|98-100|}{100}\times100\%=2\%. \]Therefore,
\[ \boxed{2\%}. \]Residuals
If the observed value is \(y\) and a model predicts \(\hat y\), one common residual convention is
\[ \boxed{e=y-\hat y}. \]A positive residual means the observation lies above the model prediction under this convention.
Interpolation and Extrapolation
Interpolation estimates within the range used to build or calibrate the model.
Extrapolation predicts outside that range. Extrapolation generally requires more caution because the model's assumptions may stop being reasonable.
Model Domains Matter
An algebraic formula may allow inputs that are physically impossible. Length, mass, area, volume, and elapsed time are often nonnegative in elementary applications.
A numerical solution must be checked against the meaning of the variable.
Example: Reject a Nonphysical Root
Suppose a model yields candidate elapsed times
\[ t=-4 \]and
\[ t=6. \]If the process begins at \(t=0\), the meaningful solution is
\[ \boxed{t=6}. \]Rearranging Applied Formulas
Scientific formulas often need to be solved for a different variable. From
\[ \rho=\frac{m}{V}, \]multiply by \(V\):
\[ \boxed{m=\rho V}. \]Example: Solve a Power Formula for the Input
If
\[ y=kx^2, \]then
\[ x^2=\frac{y}{k}. \]For a positive physical magnitude,
\[ \boxed{x=\sqrt{\frac{y}{k}}}. \]Sensitivity to Scaling
If \(y\propto x^3\), a \(10\%\) increase in \(x\) multiplies \(y\) by
\[ 1.1^3=1.331. \]Therefore, \(y\) increases by
\[ \boxed{33.1\%}. \]Inverse Sensitivity
If
\[ y\propto\frac1{x^2}, \]increasing \(x\) by \(10\%\) multiplies \(y\) by
\[ \frac1{1.1^2}\approx0.8264. \]Thus \(y\) decreases by approximately
\[ \boxed{17.36\%}. \]Checking a Formula with Units
If
\[ m=\rho V, \]with
\[ [\rho]=\frac{\text{kg}}{\text{m}^3} \]and
\[ [V]=\text{m}^3, \]then
\[ [m]=\frac{\text{kg}}{\text{m}^3}\cdot\text{m}^3=\text{kg}. \]The units are consistent.
Checking a Scaling Result
If every length of a similar object doubles, volume must be multiplied by
\[ 2^3=8. \]A result of only twice the volume would show that a linear scale factor was incorrectly applied to a three-dimensional quantity.
Checking Variation Models
For direct variation,
\[ \frac{y}{x}=k. \]For inverse variation,
\[ xy=k. \]For a power law,
\[ \frac{y}{x^n}=k. \]These invariants provide quick consistency checks across data points.
Common Error: Ignoring Units
Substituting centimeters into a formula calibrated in meters can create a large error. Convert measurements to the units expected by the model before substitution.
Common Error: Scaling Area Linearly
If linear dimensions are multiplied by \(s\), area is multiplied by
\[ \boxed{s^2}, \]not \(s\).
Common Error: Scaling Volume by the Area Factor
For similar objects, volume scales by
\[ \boxed{s^3}. \]Using \(s^2\) confuses area scaling with volume scaling.
Common Error: Assuming Every Linear Model Is Proportional
Direct proportionality must have form
\[ y=kx. \]A model \(y=mx+b\) with \(b\ne0\) is linear but not directly proportional.
Common Error: Treating the Constant as Unitless
A proportionality constant may carry units. Ignoring them can hide an inconsistent model or a missed conversion.
Common Error: Extending a Model Too Far
A model fitted over one range may fail outside that range. Extrapolation should be interpreted more cautiously than interpolation.
Common Error: Accepting Every Algebraic Root
Reject roots that violate the physical meaning of the variable, such as negative elapsed time or negative geometric length in an ordinary magnitude model.
Common Error: Confusing Fit with Mechanism
An equation can fit observed data without proving why the relationship occurs. Algebra analyzes the model, while scientific interpretation also depends on evidence, experimental design, and assumptions.
A Reliable Procedure for Scientific Formulas
- Define every variable.
- Record the units.
- Convert measurements to compatible units.
- Rearrange the formula for the requested variable.
- Substitute numerical values.
- Check dimensional consistency.
- Check physical feasibility.
- Interpret the result within the model's stated assumptions.
A Reliable Procedure for Scaling Problems
- Identify the linear scale factor \(s\).
- Identify the kind of quantity being scaled.
- Use \(s\) for lengths.
- Use \(s^2\) for areas.
- Use \(s^3\) for volumes.
- Use \(s^n\) for a general power law.
- Check whether the magnitude and direction of change are reasonable.
A Reliable Procedure for Model Calibration
- Identify the proposed functional form.
- Substitute the calibration data.
- Solve for the unknown constant or parameters.
- Write the completed model.
- Test the model against the calibration data.
- Use units to interpret the parameters.
- Distinguish interpolation from extrapolation when predicting.
Practice 1: Direct Proportion
A quantity \(y\) is directly proportional to \(x\). If \(y=18\) when \(x=6\), find the model and then find \(y\) when \(x=11\).
Practice 2: Inverse Proportion
A quantity \(T\) varies inversely with \(r\). If \(T=15\) when \(r=4\), find \(T\) when \(r=10\).
Practice 3: Joint Variation
Suppose \(y=kxz\). If \(y=60\) when \(x=3\) and \(z=4\), find \(k\) and then find \(y\) when \(x=5\) and \(z=2\).
Practice 4: Square-Law Scaling
If \(y\propto x^2\), by what factor does \(y\) change when \(x\) is multiplied by \(3\)?
Practice 5: Inverse-Square Scaling
If \(y\propto1/x^2\), by what factor does \(y\) change when \(x\) doubles?
Practice 6: Area Scaling
A similar design is enlarged by linear factor \(2.5\). By what factor does its area change?
Practice 7: Volume Scaling
For the same enlargement factor \(2.5\), by what factor does volume change?
Practice 8: Density
A sample has mass \(810\) g and volume \(300\text{ cm}^3\). Find its density.
Practice 9: Rearrange a Formula
Solve
\[ p=\frac{F}{A} \]for \(A\).
Practice 10: Calibrate a Power Law
Suppose \(y=kx^3\). If \(y=54\) when \(x=3\), find \(k\).
Practice 11: Infer a Power
Doubling \(x\) multiplies \(y\) by \(16\). If \(y=kx^n\), find \(n\).
Practice 12: Scale Model
A model is built at scale \(1:40\). A model length is \(6\) cm. Find the full-size length in meters.
Practice 13: Percent Error
A model predicts \(147\) while the reference value is \(150\). Find the absolute percent error.
Practice 14: Linear Calibration
A sensor reads \(3\) when the true input is \(0\) and \(13\) when the true input is \(5\). Assuming \(y=mx+b\), find the model.
Practice 15: Identify the Error
A cube-shaped model is enlarged so every edge is doubled. A student says the new volume is twice the original volume. Explain the error and give the correct factor.
Practice Answers
Practice 1
\[ 18=6k, \] \[ k=3. \]Therefore,
\[ \boxed{y=3x}. \]At \(x=11\),
\[ \boxed{y=33}. \]Practice 2
\[ 15=\frac{k}{4}, \] \[ k=60. \]Then
\[ T=\frac{60}{10}=6. \] \[ \boxed{6}. \]Practice 3
\[ 60=k(3)(4), \] \[ k=5. \]Then
\[ y=5(5)(2)=50. \] \[ \boxed{k=5,\qquad y=50}. \]Practice 4
\[ 3^2=9. \] \[ \boxed{\text{multiplied by }9}. \]Practice 5
\[ \frac1{2^2}=\frac14. \] \[ \boxed{\text{one-fourth as large}}. \]Practice 6
\[ 2.5^2=6.25. \] \[ \boxed{6.25}. \]Practice 7
\[ 2.5^3=15.625. \] \[ \boxed{15.625}. \]Practice 8
\[ \rho=\frac{810}{300}=2.7. \] \[ \boxed{2.7\text{ g/cm}^3}. \]Practice 9
\[ p=\frac{F}{A}. \] \[ pA=F. \] \[ \boxed{A=\frac{F}{p}}. \]Practice 10
\[ 54=k(3)^3=27k. \] \[ \boxed{k=2}. \]Practice 11
\[ 16=2^n=2^4. \] \[ \boxed{n=4}. \]Practice 12
\[ 6(40)=240\text{ cm}. \] \[ 240\text{ cm}=2.4\text{ m}. \] \[ \boxed{2.4\text{ m}}. \]Practice 13
\[ \frac{|147-150|}{150}\times100\%=2\%. \] \[ \boxed{2\%}. \]Practice 14
\[ m=\frac{13-3}{5-0}=2. \]The intercept is \(3\), so
\[ \boxed{y=2x+3}. \]Practice 15
Volume scales with the cube of the linear factor:
\[ 2^3=8. \]Therefore,
\[ \boxed{\text{the volume is }8\text{ times the original}}. \]What to Remember
Scientific and engineering models are algebraic relationships with units, domains, and assumptions.
Direct variation:
\[ \boxed{y=kx}. \]Inverse variation:
\[ \boxed{y=\frac{k}{x}}. \]Power law:
\[ \boxed{y=kx^n}. \]If \(x\) is scaled by \(s\), a power-law output scales by
\[ \boxed{s^n}. \]For similar geometric objects, lengths scale as \(s\), areas as \(s^2\), and volumes as \(s^3\).
The most important habit is to treat units, domains, and scaling exponents as part of the algebra rather than decorations added after the calculation.
Where We Go Next
Scientific and engineering models often connect several quantities at once. The next family broadens that idea to systems in which quantities move through linked paths, capacities constrain what can happen, and several algebraic conditions must hold simultaneously.
The next lesson, Network, Flow, and Constraint Systems, develops those interconnected models.