Este documento describe el método cuasi-Newton para encontrar el mínimo de una función. Explica que este método genera direcciones de búsqueda recurrentes basadas en la relación entre el gradiente y la Hessiana aproximada. Luego, provee un ejemplo numérico para minimizar la función de Rosenbrock y muestra los resultados iterativos que convergen al mínimo local.
Brevísima Intruducción a las Sumas de RiemannJuliho Castillo
Estas notas presentan una introducción informal a las sumas de Riemann, diferentes métodos para calcularlas y una introducción al uso del sistema algebraico de computo SageMath para su estudio.
Pt, Co, Fe and Ni Nan particles on Micro/Nano-Structured Carbon for the Metha...David Macias Ferrer
In this work, Pt, Co, Fe and Ni nanoparticles supported on micro/nano-structured carbon (MNC) were tested in methanol electro-oxidation process in acid medium. The samples Pt/MNC, Co/MNC, Fe/MNC and Ni/MNC (10 wt % metal loading), were synthesized by impregnation method and chemical reduction route using citric acid as dispersing agent, ammonium hydroxide and a static atmosphere of Ar-H2 as reductant agents. MNC sample was synthesized via nanocasting process with anhydrous pyrolysis at 900 °C using SBA-15 as hard template and refined sugar as carbon source. SBA-15 was prepared via sol gel using
pluronic P-123 as surfactant and tetraethoxysilane as silica precursor. The samples were characterized by means of BET, FTIR, Raman, XRD, XPS, SEM and EDS. The electrochemical measurements were carrying out by cyclic voltammetry (CV) in 0.5M H2SO4 + 1.0M CH3OH at 20 mV/s and 10 cycles.
COMPARATIVE STUDY BETWEEN REMOVERS AGENTS OF SILICON INTO THE SYNTHESIS OF MI...David Macias Ferrer
The present study deals with a comparison between HF and NaOH with different concentration, as removers agents of silicon into the synthesis of micro/nano-structured pyrolytic carbon (MNC) via nanocasting method,
using SBA-15 as hard template. SBA-15 was prepared through sol-gel using pluronic P-123 as surfactant and tetraethoxysilane as a silicon precursor. MNC was synthesized by the nanocasting method and anhydrous pyrolysis at 800 °C using SBA-15 as structure directing agent and refined sugar as a carbon precursor. Testing six samples of MNC, the silicon removal was carried out employ solutions 1M NaOH, 2M NaOH, 4M NaOH, 8M NaOH, 8.6M HF and 28.9M HF by 24 h. The prepared materials were characterized by means of N2
physisorption analysis, SEM, EDS, FTIR and Raman spectroscopy. EDS analysis revealed that 8.6M HF and 8M NaOH had the highest removal efficiency of silicon, detecting 0.21% wt Si and 0.32% wt Si respectively versus 23.8% wt Si before this process, however, using solutions of a low concentration of NaOH, a silicon weight percentage lower than 0.6% is detected. Therefore, the HF could be replaced by NaOH avoiding possible damage to the structure of MNC due to the effect of HF and contributing to the safety of researchers and the environment.
Las nuevas Tecnologías y el Aprendizaje de las MatemáticasDavid Macias Ferrer
Es evidente el acelerado desarrollo de la tecnología, se ha demostrado que en los últimos cincuenta años se han presentado las más revolucionarias innovaciones, y que inevitablemente toca y afecta a la sociedad en su conjunto. El lenguaje y la comunicación no han escapado de los cambios ya desde la antigüedad, la imprenta es causal de un cambio radical en el lenguaje escrito y ahora éste, ha revolucionado con la era de la electrónica y la computación; surgen las llamadas nuevas tecnologías de la comunicación y la información (NTCI) y por extensión los medios informáticos, y su principal instrumento: la computadora.
Electrochemical Evaluation of Pt/GMC and Pt/rGO for the Electro-oxidation of ...David Macias Ferrer
In this work, Pt/GMC and Pt/rGO electrocatalysts have been prepared by impregnation reduction method in which Pt precursor is chemically reduced by citric acid, ethanol and Ar-H2 dynamic atmosphere. Graphitic mesoporous carbon (GMC) sample was synthesized via nanocasting process with anhydrous pyrolysis at 1273 K using SBA-15 as hard template and purified sugar as carbon source. SBA-15 was prepared via sol gel using pluronic P-123 as surfactant and TEOS as silica precursor. Reduced graphene oxide (rGO) was synthesized by modified Hummers method using graphite as carbon precursor. The prepared materials were characterized by means of N2 physisorption analysis, X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), Raman spectroscopy, scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDS) and high resolution transmission electron microscopy (HRTEM). The performance of electrocatalysts for methanol oxidation reaction (MOR) was measured by cyclic voltammetry (CV), chronoamperometry (CA) and electrochemical impedance spectroscopy (EIS). The electrochemical characterization techniques revealed that the mass activity of Pt/GMC, Pt/GO and the commercial electrocatalyst Pt/C were 627, 332 and 308 mA/mgPt respectively as well as the carbon monoxide tolerance index ICO for these catalysts were 1.07, 1.04 and 0.63 respectively. Therefore, Pt/GMC and Pt/GO exhibit better electrocatalytic performance and best resistance to carbonaceous intermediates species for the electrooxidation of methanol.
Electrocatalytic Evaluation in Methanol Oxidation Process of PtCo Nanoparticl...David Macias Ferrer
The high reactivity of methanol with platinum and the excellent catalytic activity for electro-oxidation of methanol on pure Pt especially at low temperature (below 80°C), makes this metal a suitable anodic electrocatalyst in DMFC [1]. However, it is well known that there is a series of technical problems in DMFC that limit their marketing [2]. While Pt, which is generally supported on activated carbon with large surface area such as Vulcan XC-72, is the best catalyst for the electro-oxidation of methanol, it rapidly becomes poisoned because of the intermediate species formed during the oxidation of methanol, mainly CO, since CO molecules can be chemically adsorbed on the surface of Pt and block the active sites, producing a poor kinetic of anodic methanol oxidation due to CO poisoning and a low electrocatalytic activity of electrocatalysts [3-5]. Although electrocatalysts based on Pt and Pt-Ru alloy have shown a good catalytic activity for electro-oxidation of methanol, another of the limitations in the development of DMFC for commercial applications is the high cost of both noble metals [6-7]. In this study, we report a nanocasting method of low cost to prepare NC using SBA-15 as hard template and purified sugar as carbon precursor; the inclusion of Co as a second metal, in order to reduce the amount of platinum and measure its catalytic activity in methanol oxidation process
1º Caso Practico Lubricacion Rodamiento Motor 10CVCarlosAroeira1
Caso pratico análise analise de vibrações em rolamento de HVAC para resolver problema de lubrificação apresentado durante a 1ª reuniao do Vibration Institute em Lisboa em 24 de maio de 2024
Una señal analógica es una señal generada por algún tipo de fenómeno electromagnético; que es representable por una función matemática continua en la que es variable su amplitud y periodo en función del tiempo.
1. TecNM/Instituto Tecnológico de Cd. Madero
MÉTODOS INDIRECTOS
MÉTODO QUASI-NEWTON
(WILLIAM DAVIDON)
Dr. David Macias Ferrer
Centro de Investigación en Petroquímica
William C. Davidon
(1927-2013)
2. MÉTODO QUASI-NEWTON
Direcciones de Búsqueda, basada en la relación:
1
s H x xk k k
f
1. Como alternativa de cálculo (si se quiere evitar el cálculo de la inversa de H), se puede
resolver el siguiente sistema de ecuaciones:
... (6.15)
La relación recurrente es
de la forma:
1
x x sk k k k
El escalar k en este caso se
toma de la optimización
de:
Un Criterio de convergencia adecuado
puede ser:
xk
f
H x x xk k k
f ... (6.16)
x sk k
f
3. Vector Inicial x0
Optimizar f(xk + sk ), para encontrar k
Generar el vector xk+1
Encontrar el vector de dirección sk
Vector Óptimo xopt
Evaluar la Función Objetivo en xopt, es decir f(xopt )
Solución Óptima f(xopt )
1
s H x xk k k
f
|f(xk)|<
1
x x sk k k k
k = k + 1
Sí
No
MÉTODO QUASI-NEWTON
4. EJEMPLO
Encuentre el vector x que minimice la función
2 22
1 2 2 1 1, 100 1f x x x x x
Si: 0
1 2x
T
Esta función es conocida como la función de Rosenbrock
MÉTODO QUASI-NEWTON
5. EJEMPLO
Encuentre el vector x que minimice la función
2 22
1 2 2 1 1, 100 1f x x x x x
Si: 0
1 2x
T
MÉTODO QUASI-NEWTON
6. EJEMPLO
El gradiente es:
1 2
2 1
202 200 2
200
x
x x
f
x x
0 396
200
xf
Para x0:
La Hessiana es:
2
1 2 1
1
2 1200 400 400
400 200
H x
x x x
x
0 402 400
( )
400 200
H x
De aquí que:
10 0.00251256 0.00502513
( )
0.005025126 0.00505025
H x
0
443.6395xf El módulo del gradiente es: Si = 0.0001 y ya que: 0
xf
Se procede a encontrar la dirección s0 para ello:
Encuentre el vector x que minimice la función
2 22
1 2 2 1 1, 100 1f x x x x x
Si: 0
1 2x
T
MÉTODO QUASI-NEWTON
7. CONTINUACIÓN
Derivando respecto de : 6 3 2
4.1624 10 0.06124 199.97 199.97
df
d
10 0 0 0.00251256 0.00502513 396 0.0101
( )
0.005025126 0.00505025 200 0.9799
s H x xf
La dirección s0 es:
Aplicando la relación recurrente para k = 0:
1 0 0 0
x x s
De aquí que:
1
1
1
2
1 0.0101
2 0.9799
x
x
Luego entonces:1 1 0.0101
2 0.9799
x
Optimizando f():
2 22
6 4 2 3 2
100 2 0.9799 1 0.0101 1 1 0.0101
1.0406 10 2.0404 10 0.9999 199.97 104
f
MÉTODO QUASI-NEWTON
8. CONTINUACIÓN
1 2 30.9995, 4901.9799 9804.9307
6 3 2
4.1624 10 0.06124 199.97 199.97 0
Si f’() = 0 entonces:
Sustituyendo en la relación recurrente:
1 0 0 0
x x s
Tomando la raíz positiva mas pequeña: 0
0.9995
Luego entonces:
1 1 0.0101
0.9995
2 0.9799
x
1 1.0100
1.0205
x
Las tres raíces reales son:
Ahora bien, la Hessiana para este vector (y para todos los posteriores) se vuelve
definida positiva:
10 0.532358815 1.07541398
( )
1.07541398 2.177435565
H x
MÉTODO QUASI-NEWTON
9. Lo anterior corresponde a una etapa. El vector óptimo se logra en 14 etapas; la
siguiente tabla muestra los resultados:
k x1 x2 f(xk) |f(xk)|
0 0.999595 -1.0000 2.0000 104 443.6395
1 0.097455 -1.0100 1.0205 4.040294 3.8978
2 1.412991 -0.8015 0.5991 3.432175 19.4839
3 2.550827 -0.5376 0.2372 2.632401 17.5874
4 2.014159 -0.1923 -0.0019 1.572796 9.4548
5 2.446620 0.0813 -0.0288 0.969403 7.1149
10 0.015568 0.9910 0.9831 0.000183 0.4650
11 0.854706 0.9912 0.9834 0.000176 0.4577
12 0.264310 1.0006 1.0012 6.626E-07 0.0239
13 1 1.0004 1.0009 3.568E-07 0.0176
14 ---- 1.0000 1.0000 1.657E-11 0.0001
RESÚMEN
MÉTODO QUASI-NEWTON
10. Por lo tanto el vector óptimo que corresponde al extremo mínimo es:
1
1
xopt
Extremo Mínimo Local
Punto Óptimo
de la Función
2 22
1 2 2 1 1, 100 1f x x x x x
RESÚMEN
MÉTODO QUASI-NEWTON
11. BIBLIOGRAFÍA
T.F. Edgar, D.M. Himmelblau, L.S. Lasdon, “Optimization of Chemical Processes”, 2nd
Edition, New York, USA, McGraw Hill Inc., 2001